System

Abstract

A system is provided.SOLUTION: A system comprising: means for inputting a policy document; means for analyzing the input policy document using natural language processing techniques to extract key points; means for summarizing the extracted key points; means for selecting specific portions based on user interest and providing a detailed description; means for detecting changes in the policy document and updating the summary and description; and means for allowing a user to ask and answer questions through an interactive interface.SELECTED DRAWING: Figure 1

Description

[Technical Field]

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

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

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

[0004] In modern society, there are numerous policy documents that individuals and companies must comply with, and they are often written in complex, technical language. This makes it difficult for ordinary people to accurately understand these policy documents, resulting in problems such as not being able to properly recognize their own rights and obligations. It is also difficult to keep up to date with the latest information when the contents of policy documents are changed. Furthermore, there is a need for a system that can quickly provide information that addresses specific questions users have about policy documents. [Means for solving the problem]

[0005] The present invention provides a system that includes a means for inputting a policy document, a means for analyzing the input policy document using natural language processing technology and extracting key points, a means for summarizing the extracted key points, a means for selecting specific parts based on the user's interest and providing detailed explanations, a means for detecting changes to the policy document and updating the summary and explanation, and a means for the user to ask questions and have those questions answered through an interactive interface. This allows the user to quickly understand the gist of the policy document, and if more detailed information is needed, they can obtain explanations focused on that part. Even when the policy is changed, the latest information is automatically provided, allowing the user to always have accurate information. Furthermore, if the user has specific questions, they can have their questions answered immediately through the interactive interface.

[0006] A "policy document" is a formal document, such as a law, regulation, rule, contract, or privacy policy, that provides specific rules or guidelines.

[0007] "Input means" refers to an interface through which a user provides a policy document to the system, and is a device or function through which a document can be uploaded.

[0008] "Natural language processing technology" refers to technology for analyzing, understanding, and generating human language using computers, and includes algorithms and tools for automating text analysis and understanding.

[0009] "Analysis" refers to the process of examining the contents of a policy document in detail to understand its meaning and structure.

[0010] "Key Points" are sections of a policy document that should be given special attention and that provide important information or content for users.

[0011] A "summary" is a text that briefly summarizes the detailed contents of a policy document and provides only the key information.

[0012] "Customized Description" refers to detailed, easy-to-understand descriptions provided to users based on their specific interests and needs.

[0013] "Updating" refers to the process of reflecting changes to the content of policy documents when they are made, and the system automatically incorporates new information to keep summaries and explanations up to date.

[0014] An "interactive interface" is an interface that allows a user to ask a system questions and receive corresponding answers in real time.

[0015] "Question answering" refers to the ability to analyze relevant information and generate and provide appropriate answers to questions posed by users through an interactive interface. [Brief explanation of the drawings]

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

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

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

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

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

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

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

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

[0024] [First embodiment]

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

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

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

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

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

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

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

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

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

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

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

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

[0037] The system of the present invention efficiently analyzes, summarizes, customizes, explains, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[0038] System configuration and functions

[0039] 1. Document upload and text analysis

[0040] User:

[0041] Users can upload policy documents through a web interface.

[0042] Device:

[0043] The policy document uploaded by the user is sent from the terminal to the server.

[0044] server:

[0045] The server runs a natural language processing (NLP) engine to parse the policy document.

[0046] The NLP engine performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[0047] Important statements are selected and stored in a database on the server.

[0048] 2. Summary Generation

[0049] server:

[0050] A summary is generated based on important sentences from the analysis results.

[0051] The generated summary is presented in a simple format that is easy for the user to understand.

[0052] The summaries are stored in a database for quick access at all times.

[0053] 3. Customizable commentary

[0054] User:

[0055] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[0056] Device:

[0057] The entered topics and keywords are sent from the terminal to the server.

[0058] server:

[0059] The server generates a detailed explanation of the relevant parts based on the topic entered.

[0060] The customized explanations are stored in a database and provided to the user.

[0061] 4. Regular updates

[0062] server:

[0063] The system periodically checks for changes to the policy document.

[0064] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[0065] The user is notified of any updated information, so they can always keep up to date with the latest information.

[0066] 5. User Interaction

[0067] User:

[0068] A conversational interface allows users to enter questions about the policy.

[0069] For example, you could ask, "How does this policy protect my personal information?"

[0070] Device:

[0071] The user's question is sent from the terminal to the server.

[0072] server:

[0073] The server re-parses the relevant sections based on the question and generates an answer.

[0074] The generated answer is sent to the terminal and provided to the user.

[0075] Specific examples

[0076] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[0077] User:

[0078] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[0079] Device:

[0080] The document and topic of interest information are sent to a server.

[0081] server:

[0082] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[0083] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[0084] The summaries and customized descriptions are stored in a database and sent to the user's terminal.

[0085] Device:

[0086] Display a summary and customizable description to the user.

[0087] User:

[0088] Review the information provided and familiarize yourself with the policy.

[0089] The system allows users to quickly understand complex policy documents and easily access important information and details of interest. It provides up-to-date information and immediate responses to user questions, significantly improving user convenience and comprehension.

[0090] The processing flow will be explained below.

[0091] Step 1:

[0092] Users upload policy documents through a web interface.

[0093] Step 2:

[0094] The device receives the uploaded policy document and sends this data to the server.

[0095] Step 3:

[0096] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[0097] Step 4:

[0098] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[0099] Step 5:

[0100] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[0101] Step 6:

[0102] The summarized results are stored in a database on the server for quick access later.

[0103] Step 7:

[0104] Users enter specific topics or keywords through a web interface, such as "data retention period."

[0105] Step 8:

[0106] The device sends the topics and keywords entered by the user to the server.

[0107] Step 9:

[0108] Based on the received topics and keywords, the server reparses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[0109] Step 10:

[0110] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[0111] Step 11:

[0112] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[0113] Step 12:

[0114] When there is an update, the server notifies the user, who can then access the latest information.

[0115] Step 13:

[0116] Users enter questions through a conversational interface, such as "How is my personal information protected?"

[0117] Step 14:

[0118] The device sends the user's question to the server.

[0119] Step 15:

[0120] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[0121] Step 16:

[0122] The generated answer is sent from the server to the device.

[0123] Step 17:

[0124] The device receives the answer from the server and displays it to the user.

[0125] The above is the specific operation performed at each processing step in this system, which enables users to quickly understand complex policy documents and access the information they need.

[0126] Example 1

[0127] 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."

[0128] Conventional policy document analysis systems have had many challenges in quickly and accurately extracting important information from documents and providing users with summaries and detailed explanations. Furthermore, they have limited functionality for responding to document changes in real time and providing immediate answers to user questions. This situation makes it difficult for users to understand the complex content of policies, requiring significant time and effort, and delays in updating information make it difficult to stay up to date with the latest information.

[0129] 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.

[0130] In this invention, the server includes a means for inputting a document, a means for analyzing the input document using natural language processing technology and extracting important information, a means for summarizing the extracted important information, a means for selecting specific information based on a user's request and providing a detailed explanation, a means for detecting changes in the document and updating the summary and explanation, and a means for the user to ask questions and receive answers through an interactive interface. This allows users to quickly understand complex policy documents and easily access important information and details of interest. Furthermore, the server is always provided with the latest information and can immediately respond to user questions, significantly improving user convenience and understanding.

[0131] A "document" is text data such as a policy or a regulation, which describes specific rules or information.

[0132] "Means for input" refers to an interface that allows a user to provide a document to the system, and includes file uploading and the like.

[0133] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and includes grammatical analysis, keyword extraction, semantic analysis, etc.

[0134] The "means for analyzing" is a method that provides a function for analyzing the contents of a document using natural language processing technology and extracting important information.

[0135] "Important information" refers to parts of a document that are particularly meaningful or valuable to the user.

[0136] "Means of extraction" refers to the technology used to select and extract important information from the analysis results.

[0137] "Methods of summarizing" are methods for concisely summarizing content based on extracted important information.

[0138] "Means for providing" refers to the method for displaying the processing results, summary, and detailed explanation to the user.

[0139] A "means for detecting changes" is a method that has the function of monitoring updates to a document and automatically recognizing changes.

[0140] "Means for updating" means techniques for updating the summary and detailed description with new content when changes are detected.

[0141] An "interactive interface" is an interface that allows a user to interact directly with a system, including chatbots for inputting questions and instructions.

[0142] "Answering means" refers to a technique for generating and providing an appropriate response to a question from a user.

[0143] The system of the present invention is a system for efficiently analyzing, summarizing, customizing explanations, updating policy documents, and answering user questions. An embodiment of this system is described in detail below.

[0144] System configuration and functions

[0145] Document upload and text analysis

[0146] User:

[0147] Users upload policy documents using a web interface, either by dragging and dropping files into the browser interface or by clicking the file chooser button.

[0148] Device:

[0149] The uploaded policy document is sent to the server via the terminal, using an HTTP POST request to send the file contents to the server in multipart format.

[0150] server:

[0151] The server temporarily stores the received document and starts a natural language processing (NLP) engine (e.g., spaCy). The server reads the document text and performs grammatical analysis, keyword extraction, and importance evaluation of each sentence. The analysis results are evaluated, and important sentences are identified and stored in a PostgreSQL database.

[0152] Generate a summary

[0153] server:

[0154] The server generates summaries based on key sentences from the analysis results of the NLP engine. It then uses algorithms such as TextRank and BERT to convert the summaries into a simple format that is easy for users to understand. The summaries are stored in a PostgreSQL database and can be quickly provided upon user request.

[0155] Customizable commentary

[0156] User:

[0157] A user types into the web interface to get information on a particular topic or keyword of interest, for example, typing "data retention period" into a text box.

[0158] Device:

[0159] The topics and keywords entered are sent to the server as an HTTP POST request.

[0160] server:

[0161] The server receives the request and generates a detailed explanation of the relevant part based on the input topic. The explanation is generated using a language model (e.g., OpenAI GPT-3). The customized explanation is stored in a database and returned to the user's device.

[0162] Regular updates

[0163] server:

[0164] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-parsed with the NLP engine, and the summary and customization descriptions are automatically updated. Updates are notified to users in real time through a notification system.

[0165] User Interaction

[0166] User:

[0167] Users use a conversational interface (e.g., a chatbot) to enter questions about the policy, such as "How does this policy protect my personal information?"

[0168] Device:

[0169] The user's question is sent directly to the server, and the input content is constructed as an HTTP request in natural language format and sent.

[0170] server:

[0171] The server receives the question, again using an NLP engine or language model to identify the relevant section, generates an answer to the question, and returns the result to the device, where the generated answer is provided to the user in real time.

[0172] Specific examples

[0173] For example, a specific flow will be described below when a user requests detailed information about the "data storage period."

[0174] User:

[0175] Upload your privacy policy document via the web interface, enter "Data Retention Period" as the topic of interest, and submit.

[0176] Device:

[0177] Submit the document and topic of interest information to the server as an HTTP POST request.

[0178] server:

[0179] The policy document is analyzed using spaCy, key sentences are extracted, and a summary is generated. The section regarding "data retention period" is then analyzed again, and a detailed explanation is generated using the OpenAI GPT-3 model. The generated summary and customized explanation are stored in a PostgreSQL database and sent to the device as an HTTP response.

[0180] Device:

[0181] Receives the response from the server and displays a summary and customized explanation to the user.

[0182] User:

[0183] Review the information provided to deepen your understanding of the policy. Depending on the topic, you may have more detailed questions or request information on a different topic.

[0184] Prompt Sentence Examples

[0185] "Please extract key statements from the uploaded policy document, generate a summary, and provide a detailed explanation of the 'data retention period'."

[0186] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0187] Step 1:

[0188] User:

[0189] Users upload policy documents through a web interface. Specifically, they can click the file chooser button in their browser to select a local file, or they can drag and drop a file into the interface. The input is the policy document (e.g., a text file or a PDF file). The output is the uploaded file information.

[0190] Step 2:

[0191] Device:

[0192] The uploaded policy document is sent from the terminal to the server. The file contents are sent to the server in multipart format using an HTTP POST request. The input is the local file path, and the output is a successful transfer to the server.

[0193] Step 3:

[0194] server:

[0195] The server temporarily stores the received document and starts a natural language processing (NLP) engine. The server reads the file contents, performs grammatical analysis, extracts keywords, and evaluates the importance of each sentence. The input is the uploaded file contents, and the output is the analysis results (a list of important sentences and their importance). Specifically, it uses the Python-based spaCy to perform grammatical analysis, extracts important keywords, and evaluates the importance of each sentence using the TextRank algorithm.

[0196] Step 4:

[0197] server:

[0198] A summary is generated based on important sentences from the analysis results. Algorithms such as TextRank and BERT are used to generate the summary. The generated summary is stored in a PostgreSQL database. The input is a list of important sentences, and the output is the generated summary. Specifically, it selects important sentences and converts them into a simple format to create a summary.

[0199] Step 5:

[0200] User:

[0201] A user enters information about a particular topic or keyword into a web interface. The input is typing the topic or keyword (e.g., "data retention period") into a text box. The output is sending that information to a server.

[0202] Step 6:

[0203] Device:

[0204] The topics and keywords entered by the user are sent to the server as an HTTP POST request. The input is the topic or keyword entered by the user, and the output is the completion of sending the request to the server.

[0205] Step 7:

[0206] server:

[0207] The server receives a request from the user and re-analyzes the relevant document portion based on the input topic. It uses a language model (e.g., OpenAI GPT-3) to generate a detailed explanation and stores the result in a database. The input is the topic, keywords, and the analysis result of the stored document, and the output is a customized detailed explanation. Specifically, it re-analyzes the relevant text and uses a generative AI model to create appropriate and detailed information for the user.

[0208] Step 8:

[0209] server:

[0210] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-analyzed by the NLP engine and the summary and customized commentary are updated. The input is the latest version of the document, which is periodically retrieved, and the output is the updated summary and commentary. Specifically, the system manages document versions, and if any changes occur, the latest data is retrieved and re-analyzed.

[0211] Step 9:

[0212] User:

[0213] Users use a conversational interface to enter questions about the policy. The input is a question entered through an interface such as a chatbot (e.g., "How does this policy protect my personal information?"), and the output is the entered question being sent to the server.

[0214] Step 10:

[0215] Device:

[0216] The user's question is sent from the terminal to the server. The input is the question entered by the user, and the output is the completion of sending the request to the server.

[0217] Step 11:

[0218] server:

[0219] The server receives the question and again uses an NLP engine or language model to identify relevant sections. It generates an answer to the question and returns the result to the device. The input is the question content and the analysis result of the relevant document parts, and the output is the generated answer. Specifically, it re-analyzes the relevant text and generates an appropriate answer to the user's question.

[0220] (Application example 1)

[0221] 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."

[0222] In physical stores, employee training and customer support require rapid understanding and response to policy documents, but this requires employees to quickly grasp the vast amount of content. However, with conventional methods, this process is cumbersome, and it is difficult to update information or respond quickly to specific questions. To solve these issues, it is necessary to develop a system that can summarize policy documents, provide customized explanations, and provide optimal answers to user questions.

[0223] 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.

[0224] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on a user's interest and providing a detailed explanation; a means for detecting changes in the policy document and updating the summary and explanation; a means for a user to ask questions and have the questions answered through an interactive interface; and a means for providing a summary and customized explanation of the policy document as an application to be installed on a mobile device or wearable device so that store employees can quickly understand the information and receive necessary training. This enables store employees to quickly understand the policy document, stay up to date with the latest information, and obtain accurate answers to specific questions.

[0225] A "policy document" is a document that deals with important policies and procedures such as laws, regulations, and guidelines.

[0226] An "input means" is an interface or device that allows a user to provide information to a system.

[0227] "Natural language processing technology" is a technology that enables computers to understand and process the language that humans use on a daily basis.

[0228] "Key points" are parts of a policy document that have particularly important content or meaning.

[0229] A "summary" is a short, concise summary of the original content.

[0230] "User interests" refer to specific topics or themes that a user is particularly interested in.

[0231] A "detailed explanation" is a specific and detailed explanation intended to provide a deeper understanding of the specific content.

[0232] The "means for detecting changes" is a function that automatically detects changes when the contents of a policy document are updated.

[0233] An "interactive interface" is an interface that allows a user to directly interact with a system and ask questions or give instructions.

[0234] A "mobile device" is a portable electronic device such as a smartphone or tablet.

[0235] A "wearable device" is an electronic device that is worn by the user.

[0236] The following describes a mode for realizing the present invention: The present invention provides a series of configurations and means as an application for mobile devices or wearable devices that enables employees of brick-and-mortar stores to quickly understand policy documents and receive the necessary training.

[0237] First, a user (employee) uses a mobile or wearable device such as a smartphone, tablet, or smart glasses. A dedicated application is installed on this device, and the user can input a policy document through its interface. The device then sends the input policy document to the server.

[0238] When the server receives the policy document, it uses natural language processing (NLP) technology to analyze the document. The NLP technology used includes specific software components such as SpaCy and Transformers (Hugging Face). The NLP engine on the server analyzes the grammar of each sentence and extracts keywords to evaluate and extract key points. The extracted key sentences are stored in a database.

[0239] The server then generates a summary based on the extracted key sentences. This summary is provided to the user in a simple, easy-to-understand format and can be viewed on their device. If the user requests further elaboration on a specific topic or keyword, the server uses a generative AI model, such as a BERT-based question-answering model, to generate a customized explanation based on the user's interests.

[0240] Additionally, the server periodically checks for changes to the policy document and automatically updates the summary and customized description according to any detected changes. This updated information is then posted back to the user's device.

[0241] When a user has a specific question, they send it to the server through a conversational interface. The server then analyzes the policy document again based on the question and generates the best answer. This process also leverages the generative AI model, generating prompts such as:

[0242] Input document:

[0243] "Our stores adhere to the following security policies: First, protecting customer information... (omitted)"

[0244] question:

[0245] "Please explain in more detail how you protect customer information."

[0246] This allows associates to quickly understand information, stay up-to-date on the latest policies, and get instant answers to specific questions, leading to efficient and effective operations in the physical store.

[0247] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0248] Step 1:

[0249] A user inputs a policy document using a mobile or wearable device such as a smartphone, tablet, or smart glasses. The policy document uploaded by the user through the application interface is then sent from the device to the server.

[0250] Input: Policy document (text file)

[0251] Output: Data sent to the server (policy document)

[0252] Step 2:

[0253] The server analyzes the received policy document using natural language processing (NLP) techniques, specifically using SpaCy and Transformers to perform grammatical analysis and keyword extraction, evaluate the importance of each sentence, and extract key points.

[0254] Input: Policy document

[0255] Data processing / data calculation: grammatical analysis, keyword extraction, importance evaluation

[0256] Output: Extracted important sentences (text data)

[0257] Step 3:

[0258] The server generates a summary based on the extracted key sentences, provides the summary in a simple format that is easy for users to understand, and stores it in a database.

[0259] Input: Extracted important sentences

[0260] Data processing / data calculation: Execution of summary generation algorithms

[0261] Output: Summary (text data)

[0262] Step 4:

[0263] When a user wants to get more information about a particular topic or keyword, they enter it through the application interface, and this information is sent from the device to the server.

[0264] Input: Topic or keyword

[0265] Output: Data sent to the server (topics and keywords)

[0266] Step 5:

[0267] The server generates detailed explanations of relevant parts based on topics and keywords provided by the user. It uses a generative AI model (e.g., a BERT-based question-answering model) to generate customized explanations and stores them in a database.

[0268] Input: Topics, keywords, policy documents

[0269] Data processing / data calculation: Generating detailed explanations (generating responses from generative AI models)

[0270] Output: Detailed explanation (text data)

[0271] Step 6:

[0272] When a user has a specific question, they send it to the server through the interactive interface, which then analyzes the question, generates an appropriate prompt, and parses the policy document again.

[0273] Input: User question

[0274] Data processing / data calculation: prompt generation, document re-analysis

[0275] Output: Prompt statement, part of analysis result

[0276] Step 7:

[0277] The server generates the best answer to the user's question and sends it to the device, where the user can check the answer on the application.

[0278] Input: prompt statement, analysis result

[0279] Data processing / data calculation: Answer generation (response generation for generative AI models)

[0280] Output: Answer (text data)

[0281] Step 8:

[0282] The server periodically checks for changes to the policy document, and if a change is detected, it parses the new version and automatically updates the summary and customization description. The updated information is notified to the user.

[0283] Input: Modified version of the policy document

[0284] Data processing / data calculations: difference checks, reanalysis, updating summaries and commentary

[0285] Output: Updated summary and commentary (text data)

[0286] Through these steps, the program helps store associates quickly understand policy documents and receive necessary training, while also providing immediate, targeted answers to specific questions.

[0287] 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.

[0288] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents combined with an emotion engine. An embodiment of the system is described in detail below.

[0289] System configuration and functions

[0290] 1. Document upload and text analysis

[0291] User:

[0292] Users can upload policy documents through a web interface.

[0293] Device:

[0294] The policy document uploaded by the user is sent from the terminal to the server.

[0295] server:

[0296] The server runs a natural language processing (NLP) engine to parse the policy document.

[0297] The NLP engine performs grammatical analysis and keyword extraction, applies a model that evaluates the importance of each sentence, and selects the most important sentences.

[0298] Important sentences are selected and stored in a database on the server.

[0299] 2. Summary Generation

[0300] server:

[0301] A summary is generated based on important sentences from the analysis results.

[0302] The generated summary is presented in a simple format that is easy for the user to understand.

[0303] The summaries are stored in a database for quick access at all times.

[0304] 3. Customizable commentary

[0305] User:

[0306] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[0307] Device:

[0308] The entered topics and keywords are sent from the terminal to the server.

[0309] server:

[0310] The server generates a detailed explanation of the relevant parts based on the topic entered.

[0311] The customized explanations are stored in a database and provided to the user.

[0312] 4. Regular updates

[0313] server:

[0314] The system periodically checks for changes to the policy document.

[0315] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[0316] The user is notified of any updated information, so they can always keep up to date with the latest information.

[0317] 5. User Interaction and Emotion Recognition

[0318] User:

[0319] A user can enter a question through an interactive interface.

[0320] For example, you could ask, "How does this policy protect my personal information?"

[0321] Device:

[0322] Based on the user's questions and usage, the device monitors the user's emotions through an emotion engine.

[0323] The emotion engine analyzes the user's emotional state and sends the results to the server.

[0324] server:

[0325] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[0326] Specific examples

[0327] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[0328] User:

[0329] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[0330] Device:

[0331] The document and topic of interest information are sent to the server, and the user's emotional state is also monitored through the emotion engine.

[0332] server:

[0333] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[0334] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[0335] The summaries and customized commentary are stored in a database and are presented to the user in an optimized format based on the results of the sentiment engine.

[0336] Device:

[0337] A summary and customized explanation are displayed to the user. When the user enters a question in the interactive interface, the server generates an answer that takes into account the user's emotional state.

[0338] server:

[0339] Answers to questions are generated and adjusted based on the user's emotional state.

[0340] Device:

[0341] The adjusted answer is displayed to the user.

[0342] This system not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state, further improving user convenience and understanding and reducing stress.

[0343] The processing flow will be explained below.

[0344] Step 1:

[0345] Users upload policy documents through a web interface.

[0346] Step 2:

[0347] The device receives the uploaded policy document and sends this data to the server.

[0348] Step 3:

[0349] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[0350] Step 4:

[0351] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[0352] Step 5:

[0353] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[0354] Step 6:

[0355] The summarized results are stored in a database on the server for quick access later.

[0356] Step 7:

[0357] Users enter specific topics or keywords through a web interface, such as "data retention period."

[0358] Step 8:

[0359] The device sends the topics and keywords entered by the user to the server.

[0360] Step 9:

[0361] Based on the received topics and keywords, the server reparses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[0362] Step 10:

[0363] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[0364] Step 11:

[0365] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[0366] Step 12:

[0367] When there is an update, the server notifies the user, who can then access the latest information.

[0368] Step 13:

[0369] Users can enter questions through a conversational interface, such as "How is my personal information protected?"

[0370] Step 14:

[0371] The device sends the user's question to the server.

[0372] Step 15:

[0373] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[0374] Step 16:

[0375] When generating an answer, the emotion engine acquires emotion data from the user's device, allowing it to understand the emotional state of the user when they asked the question.

[0376] Step 17:

[0377] The server adjusts the tone and level of detail of the answer based on the emotional data. For example, if the user is feeling stressed, it will provide a more detailed and understandable explanation.

[0378] Step 18:

[0379] The generated answer is sent from the server to the device.

[0380] Step 19:

[0381] The device receives the answer from the server and displays it to the user.

[0382] Step 20:

[0383] Users can review the displayed answers and gain a deeper understanding of the policy content.

[0384] The above are the specific operations performed at each processing step in this system. This not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state. This further improves user convenience and understanding, and reduces stress.

[0385] Example 2

[0386] 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."

[0387] Conventional policy document analysis systems not only take time to understand complex documents and extract important information, but also have difficulty reducing user stress because they do not take into account the user's emotional state. Furthermore, when the content of a policy document is changed, updates are not automatically reflected, making it difficult to access the latest information. These problems need to be solved.

[0388] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing detailed explanations; a means for detecting changes to the policy document and updating the summaries and explanations; a means for the user to ask questions through an interactive interface and answer those questions; and a means for monitoring the user's emotional state through an emotion engine and adjusting the tone and level of detail of explanations and answers. This not only allows the user to quickly understand important information and easily access the latest information, but also enables polite responses according to the user's emotional state, thereby reducing stress for the user.

[0389] A "policy document" is a document that describes the policies, rules, and procedures of a company or organization.

[0390] "Natural language processing technology" refers to technology that allows computers to analyze, understand, and generate human language.

[0391] "Key Points" refers to information or elements within a policy document that deserve special attention.

[0392] "Extraction" refers to the process of extracting necessary information from documents or data.

[0393] An "abstract" is a brief description of the overall content of a document.

[0394] "User interests" refers to information or topics that are of particular interest or concern to a user.

[0395] "Specific portion" refers to a specific section or paragraph within a policy document.

[0396] A "detailed description" is a detailed and thorough description of a particular piece of information or topic.

[0397] "Change detection" refers to the ability to automatically identify content changes in policy documents.

[0398] "Updating" refers to the act of modifying data or descriptions to reflect new information or changes.

[0399] An "interactive interface" refers to an interface that allows a user to interact with a system to obtain information.

[0400] An "emotion engine" refers to a technology or system that analyzes a user's emotional state from their input and behavior and evaluates that state.

[0401] "Tone and detail adjustment" refers to the ability to appropriately change the wording and detail of explanations and answers provided depending on the user's emotional state.

[0402] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[0403] System configuration and functions

[0404] 1. Document upload and text analysis

[0405] User: Uploads a policy document through the web interface, or uses the file selector to select the document and clicks the upload button.

[0406] Terminal: Once a document is uploaded, it is sent to the server. It also has a function to display the progress according to the user's operations.

[0407] Server: Upon receiving the policy document, it launches a natural language processing (NLP) engine. The NLP engine performs grammatical analysis and keyword extraction, evaluates the importance of each sentence, and extracts important sentences. This extraction process uses technologies such as text analysis engines and machine learning models. Once the important sentences are extracted, they are stored in a database.

[0408] 2. Summary Generation

[0409] Server: The NLP engine generates a summary based on the extracted key sentences. The summary is formatted in a simple, easy-to-understand format and stored in the server's database, allowing users to quickly access it later.

[0410] 3. Customizable commentary

[0411] User: Enters a specific topic or keyword in the web interface. For example, the user enters an interest such as "data retention period."

[0412] Terminal: The input topic and keyword information is sent to the server, while the emotion engine monitors the user's emotional state.

[0413] Server: Based on the topic entered, the server will generate a detailed explanation of the relevant part. For example, it will re-analyze the section about "Data Retention Period" and generate a detailed explanation. This customized explanation will be stored in the database and provided to the user.

[0414] 4. Regular updates

[0415] Server: The system periodically checks for changes to the policy document, for example by checking the file modification date and time on a daily basis. If a change is detected, it automatically analyzes the new version and updates the summary and customization description.

[0416] Server: The updated information is saved in the database and notified to the user, who will see a notification message on their dashboard saying "There is an update."

[0417] 5. User Interaction and Emotion Recognition

[0418] User: Enters a question through a conversational interface, for example, a specific question such as "How does this policy protect my personal information?"

[0419] Device: Based on the user's questions and emotional state, the emotion engine monitors the user's emotions and sends this information to the server.

[0420] Server: The server analyzes the emotional data and adjusts the tone and level of detail of the explanations and responses. For example, if the user is confused or stressed, the server will provide a more detailed and understandable explanation.

[0421] Terminal: Display the adjusted answer to the user. This process is repeated each time the user enters a new question.

[0422] Specific examples

[0423] For example, a user may upload a privacy policy document and request more information about "data retention period."

[0424] User: Uploads a privacy policy document via the web interface and fills in the form field with the topic "Data Retention Period".

[0425] Terminal: Sends document and topic information to the server. The user's emotional state is also monitored through the emotion engine.

[0426] Server: Analyzes the privacy policy document, extracts key sentences, and generates a summary. For example, selects the sentence "Data will be stored for a minimum of five years" as a key sentence. Re-analyzes the section on "Data storage period" and generates a detailed explanation, for example, "Data will be stored for a minimum of five years and then deleted." The summary and customized explanation are stored in a database and provided to the user in an optimized format based on the results of the sentiment engine.

[0427] Device: Display a summary and customized explanation to the user. For example, a dashboard displays a summary and detailed explanation.

[0428] Users: Check if it contains the information they want to know.

[0429] Server: Generates answers to questions as users enter follow-up questions in the conversational interface. Based on the analysis results of the emotion engine, the server adjusts the tone of the answers. For example, if the user is confused, it provides more detailed and polite explanations.

[0430] Terminal: Display the adjusted answer to the user.

[0431] Prompt Sentence Examples

[0432] An example of a prompt to input to a generative AI model could be, "Please summarize the privacy policy document you uploaded and provide detailed information about data retention periods."

[0433] This system allows users to quickly understand complex policy documents and efficiently access the information they need, while also minimizing user stress and confusion through the use of an emotion engine.

[0434] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0435] Step 1: User uploads document

[0436] User: Uploads a policy document through a web interface. For example, the user selects a PDF file and clicks the upload button.

[0437] Input: A policy document file selected by the user.

[0438] Output: The uploaded policy document file is transferred to the device.

[0439] Specific Action: The user selects a file using the file input field in the browser and presses the upload button.

[0440] Step 2: Send the policy document to the server

[0441] Terminal: Sends the uploaded policy document to the server. Displays the sending status with a progress bar.

[0442] Input: The policy document file uploaded by the user.

[0443] Output: The policy document file transferred to the server.

[0444] Specific behavior: The device sends the uploaded file as an HTTP request to a specific API endpoint, and displays a success message when the file is successfully sent to the server.

[0445] Step 3: Server launches NLP engine and analyzes

[0446] Server: Runs a natural language processing (NLP) engine to analyze the policy document, performing grammar analysis, keyword extraction, and weighting of each sentence.

[0447] Input: The policy document file transferred to the server.

[0448] Output: A list of extracted important sentences.

[0449] What it does: The server calls the analysis engine's API to analyze the text data of the document. The engine tokenizes the text, identifies grammatical patterns, and evaluates and extracts important phrases and sentences.

[0450] Step 4: Save important statements to the database

[0451] Server: Store the extracted important sentences in a database.

[0452] Input: A list of extracted important sentences.

[0453] Output: Key statements stored in the database.

[0454] Specific operation: The server inserts the extracted important sentences into the database as structured data.

[0455] Step 5: Generate a summary

[0456] Server: Generates a summary based on important sentences from the analysis results. The generated summary is formatted in a simple format.

[0457] Input: Key statements stored in the database.

[0458] Output: The generated summary.

[0459] Specific actions: The server sorts the important sentences, creates a summary in a concise paragraph format, and stores it in the database.

[0460] Step 6: Enter a customizable description

[0461] User: Enter a specific topic or keyword in the web interface.

[0462] Input: The topic or keyword entered by the user.

[0463] Output: Topics and keywords sent to the device.

[0464] Specific actions: The user enters keywords of interest into the search box on the interface and presses the search button.

[0465] Step 7: Send topic information to the server

[0466] Terminal: The input topic and keyword information is sent to the server. At the same time, the emotion engine monitors the user's emotional state.

[0467] Input: The topic or keyword entered by the user.

[0468] Output: The topics and keywords sent to the server.

[0469] Specific operation: The device sends an HTTP request containing the user's topic information and emotional state data to the server.

[0470] Step 8: Generate customization instructions

[0471] Server: Generates detailed explanations of relevant parts based on the input topic.

[0472] Input: The topics or keywords sent to the server.

[0473] Output: The generated customization description.

[0474] Specific operation: The server searches the database based on the topic or keyword, retrieves relevant information, and generates a detailed explanation. The generated explanation is stored in the database.

[0475] Step 9: Provide customization instructions

[0476] Terminal: Display the generated customization description to the user.

[0477] Input: The generated customization description.

[0478] Output: Customization instructions displayed to the user.

[0479] Specific operation: The terminal displays the customization explanation received from the server on the interface.

[0480] Step 10: Check for changes to the policy document

[0481] Server: Periodically check for changes to the policy document.

[0482] Input: The latest policy document.

[0483] Output: The changes detected.

[0484] Specific operation: The server checks the file's update date and time and hash value to determine whether there have been any changes.

[0485] Step 11: Update the summary and description

[0486] Server: If a change is detected, the new version is automatically analyzed and the summary and customization description are updated.

[0487] Input: The most recent policy document where a change was detected.

[0488] Output: Updated summary and commentary.

[0489] What happens: The server re-parses the changes, regenerates the summary and description, and stores them in the database.

[0490] Step 12: Notify users

[0491] Server: Sends notifications of updated information to users.

[0492] Input: Updated summary and commentary.

[0493] Output: The notification sent to the user.

[0494] What happens: The server displays a notification on the user's dashboard saying "Updates available."

[0495] Step 13: User enters question

[0496] User: Enters a question through a conversational interface.

[0497] Input: The question entered by the user.

[0498] Output: The question sent to the terminal.

[0499] Specific actions: The user enters a question into the question box on the interface and presses the submit button.

[0500] Step 14: Emotional state monitoring by the emotion engine

[0501] Terminal: Sends the user's question and emotional state to the server.

[0502] Input: User-entered question and emotional state data.

[0503] Output: Questions and emotional state data sent to the server.

[0504] Specific operation: The device analyzes the user's input using the emotion engine, generates emotional state data, and sends this data and the question to the server.

[0505] Step 15: Generate and refine explanations and answers

[0506] Server: Generates answers to questions and adjusts the answers based on sentiment data.

[0507] Input: User-entered question and emotional state data.

[0508] Output: The adjusted answer.

[0509] Specific operation: The server analyzes the question content and emotional state, generates a detailed explanation to make it easier to access, and adjusts the explanation based on the emotional data.

[0510] Step 16: Providing a tailored solution

[0511] Terminal: Display the adjusted answer to the user.

[0512] Input: The adjusted answer.

[0513] Output: The adjusted answer displayed in the interface.

[0514] Specific operation: The device displays the adjusted answer on the interface, allowing the user to confirm the answer.

[0515] The above are the processing steps of this system and their specific operations.

[0516] (Application example 2)

[0517] 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."

[0518] In autonomous vehicles, instruction manuals and operating instructions are complex, making it difficult for users to quickly and accurately obtain information. This often leads to confusion and stress. Furthermore, operating errors in such situations could lead to serious accidents. Therefore, there is a need for real-time, customized explanations and question responses that take users' emotions into consideration.

[0519] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0520] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing a detailed explanation; a means for detecting changes to the policy document and updating the summary and explanation; a means for the user to ask questions through an interactive interface and have the questions answered; a means for analyzing the document, extracting important sentences, and generating a summary; and a means for analyzing the user's emotions using an emotion engine and providing a detailed explanation according to the user's emotional state. This allows the user to quickly and accurately understand the instruction manual or operating manual for the autonomous vehicle and receive detailed explanations according to their emotional state, reducing confusion and stress and enabling safe and effective operation.

[0521] "Policy Document" refers to a document regarding various policies and regulations provided to users.

[0522] An "input means" is a device or interface for sending a policy document to a server.

[0523] "Natural language processing technology" is a computer technology for analyzing text data and understanding its meaning.

[0524] "Key Points" are information that is considered particularly important in a policy document.

[0525] An "extraction means" is a mechanism for selecting information from within a document based on certain criteria.

[0526] A "summarizing tool" is a method for concisely summarizing the extracted important points.

[0527] "User interests" are information that a user is particularly interested in or wants to know about.

[0528] "Detailed Description" refers to a detailed explanation of information provided based on the user's interest.

[0529] "Means for detecting changes" refers to the ability to automatically recognize changes to policy documents when they occur.

[0530] An "interactive interface" is a user interface that allows users to directly input questions and instructions into the system.

[0531] An "emotion engine" is a system for analyzing and understanding a user's emotional state.

[0532] "Emotional state" refers to the state of mind that a user feels in a particular situation.

[0533] The system of the present invention is designed to analyze, summarize, customize, and update policy documents, and to answer user questions. In this embodiment, the system is particularly applied to instruction manuals and operation manuals for autonomous vehicles.

[0534] System configuration and functions

[0535] 1. Document upload and text analysis

[0536] Users upload policy documents through input means such as smartphones or smart glasses.

[0537] The terminal transmits the uploaded policy document to the server.

[0538] The server uses natural language processing techniques to parse the policy document and extract key points, including grammatical analysis, keyword extraction, and weighting each sentence.

[0539] The extracted important sentences are stored in a database.

[0540] 2. Summary Generation

[0541] The server generates a summary based on key sentences from the analysis results and provides it in a simple format that is easy for users to understand.

[0542] The generated summaries are stored in a database for quick access.

[0543] 3. Customizable commentary

[0544] Users enter topics or keywords of interest into an interactive interface.

[0545] The terminal sends this to the server.

[0546] Based on the topic entered, the server generates a detailed explanation of the relevant parts, including information on specific features and procedures.

[0547] 4. Regular updates

[0548] The server periodically checks for changes to the policy document and automatically parses the new version to update the summary and customization description.

[0549] The updated information will be notified to the user.

[0550] 5. User Interaction and Emotion Recognition

[0551] A user inputs a question through an interactive interface.

[0552] Example: "What are the conditions for airbag deployment?"

[0553] The device monitors the user's emotions through an emotion engine, which analyzes the user's emotional state and sends the results to the server.

[0554] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[0555] Specific examples

[0556] For example, if a user uploads a driverless vehicle instruction manual to their smart glasses and requests detailed information about "airbag deployment conditions," if the emotion engine recognizes "confusion," it will provide a polite response like this:

[0557] Example prompt sentence:

[0558] What are the conditions for airbag deployment?

[0559] Example answer:

[0560] Regarding the conditions for airbag deployment, the airbag system must first detect an emergency situation. It will deploy at a specific speed and angle of impact. Shall I explain this in more detail?

[0561] In this way, the emotion engine adjusts responses according to the user's emotional state, allowing the user to obtain more appropriate and reliable information, preventing operational errors and improving safety.

[0562] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0563] Step 1:

[0564] Users upload policy documents using their smartphones or smart glasses.

[0565] Input: Policy document

[0566] Output: The policy document is sent to the server.

[0567] Specific operation: The user selects a file through the device interface and clicks the upload button.

[0568] Step 2:

[0569] The terminal sends the uploaded policy document to the server.

[0570] Input: A user-uploaded policy document

[0571] Output: The policy document is transferred to the server.

[0572] Specific operation: By clicking the upload button, the device sends an HTTP request to the server.

[0573] Step 3:

[0574] The server uses natural language processing techniques to parse the policy document and extract key points.

[0575] Input: The policy document sent to the server

[0576] Output: Extracted key points

[0577] Specific operation: The server passes the policy document to an analysis tool, which performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[0578] Step 4:

[0579] The server generates a summary of the extracted key points.

[0580] Input: Key points extracted from the analysis

[0581] Output: Generated summary

[0582] What it does: Select the most important sentences and combine them to form a concise summary.

[0583] Step 5:

[0584] Users enter topics or keywords of interest into an interactive interface.

[0585] Input: Topics or keywords entered by the user

[0586] Output: Topics and keywords sent to the server

[0587] Specific actions: The user enters an item of interest into a text box on the interface and clicks the submit button.

[0588] Step 6:

[0589] The device sends the entered topics and keywords to the server.

[0590] Input: Topic or keyword

[0591] Output: Topics and keywords are sent to the server

[0592] Specific operation: By clicking the send button, the device sends an HTTP request to the server.

[0593] Step 7:

[0594] The server generates a detailed explanation of the relevant parts based on the topic entered.

[0595] Input: Topics or keywords sent to the server

[0596] Output: Detailed explanation generated

[0597] What happens: The server uses natural language processing techniques to re-parse the relevant section and generate a detailed explanation.

[0598] Step 8:

[0599] The server uses an emotion engine to analyze the user's emotions and provide detailed explanations according to their emotional state.

[0600] Input: User emotion data

[0601] Output: Detailed explanation according to emotional state

[0602] Specific operation: The emotion engine analyzes the user's emotional data and adjusts the content and tone of the commentary based on the results.

[0603] Step 9:

[0604] The server periodically checks for changes to the policy document and updates the summary and customization description.

[0605] Input: New policy document

[0606] Output: Updated summary and customization commentary

[0607] What it does: The server periodically reparses the document and updates the summary and detailed description if it detects any changes.

[0608] Step 10:

[0609] A user enters a question through an interactive interface and receives an answer to that question.

[0610] Input: User question

[0611] Output: The answer provided to the user

[0612] Specific operation: The user enters a question into a text box on the interface, clicks the submit button, and the server generates and displays the answer.

[0613] 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.

[0614] 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.

[0615] 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.

[0616] [Second embodiment]

[0617] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0618] 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.

[0619] 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).

[0620] 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.

[0621] 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.

[0622] 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).

[0623] 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.

[0624] 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.

[0625] 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.

[0626] 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.

[0627] 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.

[0628] 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."

[0629] The system of the present invention efficiently analyzes, summarizes, customizes, explains, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[0630] System configuration and functions

[0631] 1. Document upload and text analysis

[0632] User:

[0633] Users can upload policy documents through a web interface.

[0634] Device:

[0635] The policy document uploaded by the user is sent from the terminal to the server.

[0636] server:

[0637] The server runs a natural language processing (NLP) engine to parse the policy document.

[0638] The NLP engine performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[0639] Important statements are selected and stored in a database on the server.

[0640] 2. Summary Generation

[0641] server:

[0642] A summary is generated based on important sentences from the analysis results.

[0643] The generated summary is presented in a simple format that is easy for the user to understand.

[0644] The summaries are stored in a database for quick access at all times.

[0645] 3. Customizable commentary

[0646] User:

[0647] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[0648] Device:

[0649] The entered topics and keywords are sent from the terminal to the server.

[0650] server:

[0651] The server generates a detailed explanation of the relevant parts based on the topic entered.

[0652] The customized explanations are stored in a database and provided to the user.

[0653] 4. Regular updates

[0654] server:

[0655] The system periodically checks for changes to the policy document.

[0656] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[0657] The user is notified of any updated information, so they can always keep up to date with the latest information.

[0658] 5. User Interaction

[0659] User:

[0660] A conversational interface allows users to enter questions about the policy.

[0661] For example, you could ask, "How does this policy protect my personal information?"

[0662] Device:

[0663] The user's question is sent from the terminal to the server.

[0664] server:

[0665] The server re-parses the relevant sections based on the question and generates an answer.

[0666] The generated answer is sent to the terminal and provided to the user.

[0667] Specific examples

[0668] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[0669] User:

[0670] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[0671] Device:

[0672] The document and topic of interest information are sent to a server.

[0673] server:

[0674] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[0675] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[0676] The summaries and customized descriptions are stored in a database and sent to the user's terminal.

[0677] Device:

[0678] Display a summary and customizable description to the user.

[0679] User:

[0680] Review the information provided and familiarize yourself with the policy.

[0681] The system allows users to quickly understand complex policy documents and easily access important information and details of interest. It provides up-to-date information and immediate responses to user questions, significantly improving user convenience and comprehension.

[0682] The processing flow will be explained below.

[0683] Step 1:

[0684] Users upload policy documents through a web interface.

[0685] Step 2:

[0686] The device receives the uploaded policy document and sends this data to the server.

[0687] Step 3:

[0688] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[0689] Step 4:

[0690] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[0691] Step 5:

[0692] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[0693] Step 6:

[0694] The summarized results are stored in a database on the server for quick access later.

[0695] Step 7:

[0696] Users enter specific topics or keywords through a web interface, such as "data retention period."

[0697] Step 8:

[0698] The device sends the topics and keywords entered by the user to the server.

[0699] Step 9:

[0700] Based on the received topics and keywords, the server reparses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[0701] Step 10:

[0702] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[0703] Step 11:

[0704] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[0705] Step 12:

[0706] When there is an update, the server notifies the user, who can then access the latest information.

[0707] Step 13:

[0708] Users enter questions through a conversational interface, such as "How is my personal information protected?"

[0709] Step 14:

[0710] The device sends the user's question to the server.

[0711] Step 15:

[0712] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[0713] Step 16:

[0714] The generated answer is sent from the server to the device.

[0715] Step 17:

[0716] The device receives the answer from the server and displays it to the user.

[0717] The above is the specific operation performed at each processing step in this system, which enables users to quickly understand complex policy documents and access the information they need.

[0718] Example 1

[0719] 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."

[0720] Conventional policy document analysis systems have had many challenges in quickly and accurately extracting important information from documents and providing users with summaries and detailed explanations. Furthermore, they have limited functionality for responding to document changes in real time and providing immediate answers to user questions. This situation makes it difficult for users to understand the complex content of policies, requiring significant time and effort, and delays in updating information make it difficult to stay up to date with the latest information.

[0721] 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.

[0722] In this invention, the server includes a means for inputting a document, a means for analyzing the input document using natural language processing technology and extracting important information, a means for summarizing the extracted important information, a means for selecting specific information based on a user's request and providing a detailed explanation, a means for detecting changes in the document and updating the summary and explanation, and a means for the user to ask questions and receive answers through an interactive interface. This allows users to quickly understand complex policy documents and easily access important information and details of interest. Furthermore, the server is always provided with the latest information and can immediately respond to user questions, significantly improving user convenience and understanding.

[0723] A "document" is text data such as a policy or a regulation, which describes specific rules or information.

[0724] "Means for input" refers to an interface that allows a user to provide a document to the system, and includes file uploading and the like.

[0725] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and includes grammatical analysis, keyword extraction, semantic analysis, etc.

[0726] The "means for analyzing" is a method that provides a function for analyzing the contents of a document using natural language processing technology and extracting important information.

[0727] "Important information" refers to parts of a document that are particularly meaningful or valuable to the user.

[0728] "Means of extraction" refers to the technology used to select and extract important information from the analysis results.

[0729] "Methods of summarizing" are methods for concisely summarizing content based on extracted important information.

[0730] "Means for providing" refers to the method for displaying the processing results, summary, and detailed explanation to the user.

[0731] A "means for detecting changes" is a method that has the function of monitoring updates to a document and automatically recognizing changes.

[0732] "Means for updating" means techniques for updating the summary and detailed description with new content when changes are detected.

[0733] An "interactive interface" is an interface that allows a user to interact directly with a system, including chatbots for inputting questions and instructions.

[0734] "Answering means" refers to a technique for generating and providing an appropriate response to a question from a user.

[0735] The system of the present invention is a system for efficiently analyzing, summarizing, customizing explanations, updating policy documents, and answering user questions. An embodiment of this system is described in detail below.

[0736] System configuration and functions

[0737] Document upload and text analysis

[0738] User:

[0739] Users upload policy documents using a web interface, either by dragging and dropping files into the browser interface or by clicking the file chooser button.

[0740] Device:

[0741] The uploaded policy document is sent to the server via the terminal, using an HTTP POST request to send the file contents to the server in multipart format.

[0742] server:

[0743] The server temporarily stores the received document and starts a natural language processing (NLP) engine (e.g., spaCy). The server reads the document text and performs grammatical analysis, keyword extraction, and importance evaluation of each sentence. The analysis results are evaluated, and important sentences are identified and stored in a PostgreSQL database.

[0744] Generate a summary

[0745] server:

[0746] The server generates summaries based on key sentences from the analysis results of the NLP engine. It then uses algorithms such as TextRank and BERT to convert the summaries into a simple format that is easy for users to understand. The summaries are stored in a PostgreSQL database and can be quickly provided upon user request.

[0747] Customizable commentary

[0748] User:

[0749] A user types into the web interface to get information on a particular topic or keyword of interest, for example, typing "data retention period" into a text box.

[0750] Device:

[0751] The topics and keywords entered are sent to the server as an HTTP POST request.

[0752] server:

[0753] The server receives the request and generates a detailed explanation of the relevant part based on the input topic. The explanation is generated using a language model (e.g., OpenAI GPT-3). The customized explanation is stored in a database and returned to the user's device.

[0754] Regular updates

[0755] server:

[0756] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-parsed with the NLP engine, and the summary and customization descriptions are automatically updated. Updates are notified to users in real time through a notification system.

[0757] User Interaction

[0758] User:

[0759] Users use a conversational interface (e.g., a chatbot) to enter questions about the policy, such as "How does this policy protect my personal information?"

[0760] Device:

[0761] The user's question is sent directly to the server, and the input content is constructed as an HTTP request in natural language format and sent.

[0762] server:

[0763] The server receives the question, again using an NLP engine or language model to identify the relevant section, generates an answer to the question, and returns the result to the device, where the generated answer is provided to the user in real time.

[0764] Specific examples

[0765] For example, a specific flow will be described below when a user requests detailed information about the "data storage period."

[0766] User:

[0767] Upload your privacy policy document via the web interface, enter "Data Retention Period" as the topic of interest, and submit.

[0768] Device:

[0769] Submit the document and topic of interest information to the server as an HTTP POST request.

[0770] server:

[0771] The policy document is analyzed using spaCy, key sentences are extracted, and a summary is generated. The section regarding "data retention period" is then analyzed again, and a detailed explanation is generated using the OpenAI GPT-3 model. The generated summary and customized explanation are stored in a PostgreSQL database and sent to the device as an HTTP response.

[0772] Device:

[0773] Receives the response from the server and displays a summary and customized explanation to the user.

[0774] User:

[0775] Review the information provided to deepen your understanding of the policy. Depending on the topic, you may have more detailed questions or request information on a different topic.

[0776] Prompt Sentence Examples

[0777] "Please extract key statements from the uploaded policy document, generate a summary, and provide a detailed explanation of the 'data retention period'."

[0778] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0779] Step 1:

[0780] User:

[0781] Users upload policy documents through a web interface. Specifically, they can click the file chooser button in their browser to select a local file, or they can drag and drop a file into the interface. The input is the policy document (e.g., a text file or a PDF file). The output is the uploaded file information.

[0782] Step 2:

[0783] Device:

[0784] The uploaded policy document is sent from the terminal to the server. The file contents are sent to the server in multipart format using an HTTP POST request. The input is the local file path, and the output is a successful transfer to the server.

[0785] Step 3:

[0786] server:

[0787] The server temporarily stores the received document and starts a natural language processing (NLP) engine. The server reads the file contents, performs grammatical analysis, extracts keywords, and evaluates the importance of each sentence. The input is the uploaded file contents, and the output is the analysis results (a list of important sentences and their importance). Specifically, it uses the Python-based spaCy to perform grammatical analysis, extracts important keywords, and evaluates the importance of each sentence using the TextRank algorithm.

[0788] Step 4:

[0789] server:

[0790] A summary is generated based on important sentences from the analysis results. Algorithms such as TextRank and BERT are used to generate the summary. The generated summary is stored in a PostgreSQL database. The input is a list of important sentences, and the output is the generated summary. Specifically, it selects important sentences and converts them into a simple format to create a summary.

[0791] Step 5:

[0792] User:

[0793] A user enters information about a particular topic or keyword into a web interface. The input is typing the topic or keyword (e.g., "data retention period") into a text box. The output is sending that information to a server.

[0794] Step 6:

[0795] Device:

[0796] The topics and keywords entered by the user are sent to the server as an HTTP POST request. The input is the topic or keyword entered by the user, and the output is the completion of sending the request to the server.

[0797] Step 7:

[0798] server:

[0799] The server receives a request from the user and re-analyzes the relevant document portion based on the input topic. It uses a language model (e.g., OpenAI GPT-3) to generate a detailed explanation and stores the result in a database. The input is the topic, keywords, and the analysis result of the stored document, and the output is a customized detailed explanation. Specifically, it re-analyzes the relevant text and uses a generative AI model to create appropriate and detailed information for the user.

[0800] Step 8:

[0801] server:

[0802] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-analyzed by the NLP engine and the summary and customized commentary are updated. The input is the latest version of the document, which is periodically retrieved, and the output is the updated summary and commentary. Specifically, the system manages document versions, and if any changes occur, the latest data is retrieved and re-analyzed.

[0803] Step 9:

[0804] User:

[0805] Users use a conversational interface to enter questions about the policy. The input is a question entered through an interface such as a chatbot (e.g., "How does this policy protect my personal information?"), and the output is the entered question being sent to the server.

[0806] Step 10:

[0807] Device:

[0808] The user's question is sent from the terminal to the server. The input is the question entered by the user, and the output is the completion of sending the request to the server.

[0809] Step 11:

[0810] server:

[0811] The server receives the question and again uses an NLP engine or language model to identify relevant sections. It generates an answer to the question and returns the result to the device. The input is the question content and the analysis result of the relevant document parts, and the output is the generated answer. Specifically, it re-analyzes the relevant text and generates an appropriate answer to the user's question.

[0812] (Application example 1)

[0813] 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."

[0814] In physical stores, employee training and customer support require rapid understanding and response to policy documents, but this requires employees to quickly grasp the vast amount of content. However, with conventional methods, this process is cumbersome, and it is difficult to update information or respond quickly to specific questions. To solve these issues, it is necessary to develop a system that can summarize policy documents, provide customized explanations, and provide optimal answers to user questions.

[0815] 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.

[0816] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on a user's interest and providing a detailed explanation; a means for detecting changes in the policy document and updating the summary and explanation; a means for a user to ask questions and have the questions answered through an interactive interface; and a means for providing a summary and customized explanation of the policy document as an application to be installed on a mobile device or wearable device so that store employees can quickly understand the information and receive necessary training. This enables store employees to quickly understand the policy document, stay up to date with the latest information, and obtain accurate answers to specific questions.

[0817] A "policy document" is a document that deals with important policies and procedures such as laws, regulations, and guidelines.

[0818] An "input means" is an interface or device that allows a user to provide information to a system.

[0819] "Natural language processing technology" is a technology that enables computers to understand and process the language that humans use on a daily basis.

[0820] "Key points" are parts of a policy document that have particularly important content or meaning.

[0821] A "summary" is a short, concise summary of the original content.

[0822] "User interests" refer to specific topics or themes that a user is particularly interested in.

[0823] A "detailed explanation" is a specific and detailed explanation intended to provide a deeper understanding of the specific content.

[0824] The "means for detecting changes" is a function that automatically detects changes when the contents of a policy document are updated.

[0825] An "interactive interface" is an interface that allows a user to directly interact with a system and ask questions or give instructions.

[0826] A "mobile device" is a portable electronic device such as a smartphone or tablet.

[0827] A "wearable device" is an electronic device that is worn by the user.

[0828] The following describes a mode for realizing the present invention: The present invention provides a series of configurations and means as an application for mobile devices or wearable devices that enables employees of brick-and-mortar stores to quickly understand policy documents and receive the necessary training.

[0829] First, a user (employee) uses a mobile or wearable device such as a smartphone, tablet, or smart glasses. A dedicated application is installed on this device, and the user can input a policy document through its interface. The device then sends the input policy document to the server.

[0830] When the server receives the policy document, it uses natural language processing (NLP) technology to analyze the document. The NLP technology used includes specific software components such as SpaCy and Transformers (Hugging Face). The NLP engine on the server analyzes the grammar of each sentence and extracts keywords to evaluate and extract key points. The extracted key sentences are stored in a database.

[0831] The server then generates a summary based on the extracted key sentences. This summary is provided to the user in a simple, easy-to-understand format and can be viewed on their device. If the user requests further elaboration on a specific topic or keyword, the server uses a generative AI model, such as a BERT-based question-answering model, to generate a customized explanation based on the user's interests.

[0832] Additionally, the server periodically checks for changes to the policy document and automatically updates the summary and customized description according to any detected changes. This updated information is then posted back to the user's device.

[0833] When a user has a specific question, they send it to the server through a conversational interface. The server then analyzes the policy document again based on the question and generates the best answer. This process also leverages the generative AI model, generating prompts such as:

[0834] Input document:

[0835] "Our stores adhere to the following security policies: First, protecting customer information... (omitted)"

[0836] question:

[0837] "Please explain in more detail how you protect customer information."

[0838] This allows associates to quickly understand information, stay up-to-date on the latest policies, and get instant answers to specific questions, leading to efficient and effective operations in the physical store.

[0839] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0840] Step 1:

[0841] A user inputs a policy document using a mobile or wearable device such as a smartphone, tablet, or smart glasses. The policy document uploaded by the user through the application interface is then sent from the device to the server.

[0842] Input: Policy document (text file)

[0843] Output: Data sent to the server (policy document)

[0844] Step 2:

[0845] The server analyzes the received policy document using natural language processing (NLP) techniques, specifically using SpaCy and Transformers to perform grammatical analysis and keyword extraction, evaluate the importance of each sentence, and extract key points.

[0846] Input: Policy document

[0847] Data processing / data calculation: grammatical analysis, keyword extraction, importance evaluation

[0848] Output: Extracted important sentences (text data)

[0849] Step 3:

[0850] The server generates a summary based on the extracted key sentences, provides the summary in a simple format that is easy for users to understand, and stores it in a database.

[0851] Input: Extracted important sentences

[0852] Data processing / data calculation: Execution of summary generation algorithms

[0853] Output: Summary (text data)

[0854] Step 4:

[0855] When a user wants to get more information about a particular topic or keyword, they enter it through the application interface, and this information is sent from the device to the server.

[0856] Input: Topic or keyword

[0857] Output: Data sent to the server (topics and keywords)

[0858] Step 5:

[0859] The server generates detailed explanations of relevant parts based on topics and keywords provided by the user. It uses a generative AI model (e.g., a BERT-based question-answering model) to generate customized explanations and stores them in a database.

[0860] Input: Topics, keywords, policy documents

[0861] Data processing / data calculation: Generating detailed explanations (generating responses from generative AI models)

[0862] Output: Detailed explanation (text data)

[0863] Step 6:

[0864] When a user has a specific question, they send it to the server through the interactive interface, which then analyzes the question, generates an appropriate prompt, and parses the policy document again.

[0865] Input: User question

[0866] Data processing / data calculation: prompt generation, document re-analysis

[0867] Output: Prompt statement, part of analysis result

[0868] Step 7:

[0869] The server generates the best answer to the user's question and sends it to the device, where the user can check the answer on the application.

[0870] Input: prompt statement, analysis result

[0871] Data processing / data calculation: Answer generation (response generation for generative AI models)

[0872] Output: Answer (text data)

[0873] Step 8:

[0874] The server periodically checks for changes to the policy document, and if a change is detected, it parses the new version and automatically updates the summary and customization description. The updated information is notified to the user.

[0875] Input: Modified version of the policy document

[0876] Data processing / data calculations: difference checks, reanalysis, updating summaries and commentary

[0877] Output: Updated summary and commentary (text data)

[0878] Through these steps, the program helps store associates quickly understand policy documents and receive necessary training, while also providing immediate, targeted answers to specific questions.

[0879] 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.

[0880] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents combined with an emotion engine. An embodiment of the system is described in detail below.

[0881] System configuration and functions

[0882] 1. Document upload and text analysis

[0883] User:

[0884] Users can upload policy documents through a web interface.

[0885] Device:

[0886] The policy document uploaded by the user is sent from the terminal to the server.

[0887] server:

[0888] The server runs a natural language processing (NLP) engine to parse the policy document.

[0889] The NLP engine performs grammatical analysis and keyword extraction, applies a model that evaluates the importance of each sentence, and selects the most important sentences.

[0890] Important sentences are selected and stored in a database on the server.

[0891] 2. Summary Generation

[0892] server:

[0893] A summary is generated based on important sentences from the analysis results.

[0894] The generated summary is presented in a simple format that is easy for the user to understand.

[0895] The summaries are stored in a database for quick access at all times.

[0896] 3. Customizable commentary

[0897] User:

[0898] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[0899] Device:

[0900] The entered topics and keywords are sent from the terminal to the server.

[0901] server:

[0902] The server generates a detailed explanation of the relevant parts based on the topic entered.

[0903] The customized explanations are stored in a database and provided to the user.

[0904] 4. Regular updates

[0905] server:

[0906] The system periodically checks for changes to the policy document.

[0907] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[0908] The user is notified of any updated information, so they can always keep up to date with the latest information.

[0909] 5. User Interaction and Emotion Recognition

[0910] User:

[0911] A user can enter a question through an interactive interface.

[0912] For example, you could ask, "How does this policy protect my personal information?"

[0913] Device:

[0914] Based on the user's questions and usage, the device monitors the user's emotions through an emotion engine.

[0915] The emotion engine analyzes the user's emotional state and sends the results to the server.

[0916] server:

[0917] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[0918] Specific examples

[0919] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[0920] User:

[0921] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[0922] Device:

[0923] The document and topic of interest information are sent to the server, and the user's emotional state is also monitored through the emotion engine.

[0924] server:

[0925] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[0926] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[0927] The summaries and customized commentary are stored in a database and are presented to the user in an optimized format based on the results of the sentiment engine.

[0928] Device:

[0929] A summary and customized explanation are displayed to the user. When the user enters a question in the interactive interface, the server generates an answer that takes into account the user's emotional state.

[0930] server:

[0931] Answers to questions are generated and adjusted based on the user's emotional state.

[0932] Device:

[0933] The adjusted answer is displayed to the user.

[0934] This system not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state, further improving user convenience and understanding and reducing stress.

[0935] The processing flow will be explained below.

[0936] Step 1:

[0937] Users upload policy documents through a web interface.

[0938] Step 2:

[0939] The device receives the uploaded policy document and sends this data to the server.

[0940] Step 3:

[0941] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[0942] Step 4:

[0943] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[0944] Step 5:

[0945] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[0946] Step 6:

[0947] The summarized results are stored in a database on the server for quick access later.

[0948] Step 7:

[0949] Users enter specific topics or keywords through a web interface, such as "data retention period."

[0950] Step 8:

[0951] The device sends the topics and keywords entered by the user to the server.

[0952] Step 9:

[0953] Based on the received topics and keywords, the server reparses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[0954] Step 10:

[0955] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[0956] Step 11:

[0957] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[0958] Step 12:

[0959] When there is an update, the server notifies the user, who can then access the latest information.

[0960] Step 13:

[0961] Users can enter questions through a conversational interface, such as "How is my personal information protected?"

[0962] Step 14:

[0963] The device sends the user's question to the server.

[0964] Step 15:

[0965] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[0966] Step 16:

[0967] When generating an answer, the emotion engine acquires emotion data from the user's device, allowing it to understand the emotional state of the user when they asked the question.

[0968] Step 17:

[0969] The server adjusts the tone and level of detail of the answer based on the emotional data. For example, if the user is feeling stressed, it will provide a more detailed and understandable explanation.

[0970] Step 18:

[0971] The generated answer is sent from the server to the device.

[0972] Step 19:

[0973] The device receives the answer from the server and displays it to the user.

[0974] Step 20:

[0975] Users can review the displayed answers and gain a deeper understanding of the policy content.

[0976] The above are the specific operations performed at each processing step in this system. This not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state. This further improves user convenience and understanding, and reduces stress.

[0977] Example 2

[0978] 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."

[0979] Conventional policy document analysis systems not only take time to understand complex documents and extract important information, but also have difficulty reducing user stress because they do not take into account the user's emotional state. Furthermore, when the content of a policy document is changed, updates are not automatically reflected, making it difficult to access the latest information. These problems need to be solved.

[0980] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing detailed explanations; a means for detecting changes to the policy document and updating the summaries and explanations; a means for the user to ask questions through an interactive interface and answer those questions; and a means for monitoring the user's emotional state through an emotion engine and adjusting the tone and level of detail of explanations and answers. This not only allows the user to quickly understand important information and easily access the latest information, but also enables polite responses according to the user's emotional state, thereby reducing stress for the user.

[0981] A "policy document" is a document that describes the policies, rules, and procedures of a company or organization.

[0982] "Natural language processing technology" refers to technology that allows computers to analyze, understand, and generate human language.

[0983] "Key Points" refers to information or elements within a policy document that deserve special attention.

[0984] "Extraction" refers to the process of extracting necessary information from documents or data.

[0985] An "abstract" is a brief description of the overall content of a document.

[0986] "User interests" refers to information or topics that are of particular interest or concern to a user.

[0987] "Specific portion" refers to a specific section or paragraph within a policy document.

[0988] A "detailed description" is a detailed and thorough description of a particular piece of information or topic.

[0989] "Change detection" refers to the ability to automatically identify content changes in policy documents.

[0990] "Updating" refers to the act of modifying data or descriptions to reflect new information or changes.

[0991] An "interactive interface" refers to an interface that allows a user to interact with a system to obtain information.

[0992] An "emotion engine" refers to a technology or system that analyzes a user's emotional state from their input and behavior and evaluates that state.

[0993] "Tone and detail adjustment" refers to the ability to appropriately change the wording and detail of explanations and answers provided depending on the user's emotional state.

[0994] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[0995] System configuration and functions

[0996] 1. Document upload and text analysis

[0997] User: Uploads a policy document through the web interface, or uses the file selector to select the document and clicks the upload button.

[0998] Terminal: Once a document is uploaded, it is sent to the server. It also has a function to display the progress according to the user's operations.

[0999] Server: Upon receiving the policy document, it launches a natural language processing (NLP) engine. The NLP engine performs grammatical analysis and keyword extraction, evaluates the importance of each sentence, and extracts important sentences. This extraction process uses technologies such as text analysis engines and machine learning models. Once the important sentences are extracted, they are stored in a database.

[1000] 2. Summary Generation

[1001] Server: The NLP engine generates a summary based on the extracted key sentences. The summary is formatted in a simple, easy-to-understand format and stored in the server's database, allowing users to quickly access it later.

[1002] 3. Customizable commentary

[1003] User: Enters a specific topic or keyword in the web interface. For example, the user enters an interest such as "data retention period."

[1004] Terminal: The input topic and keyword information is sent to the server, while the emotion engine monitors the user's emotional state.

[1005] Server: Based on the topic entered, the server will generate a detailed explanation of the relevant part. For example, it will re-analyze the section about "Data Retention Period" and generate a detailed explanation. This customized explanation will be stored in the database and provided to the user.

[1006] 4. Regular updates

[1007] Server: The system periodically checks for changes to the policy document, for example by checking the file modification date and time on a daily basis. If a change is detected, it automatically analyzes the new version and updates the summary and customization description.

[1008] Server: The updated information is saved in the database and notified to the user, who will see a notification message on their dashboard saying "There is an update."

[1009] 5. User Interaction and Emotion Recognition

[1010] User: Enters a question through a conversational interface, for example, a specific question such as "How does this policy protect my personal information?"

[1011] Device: Based on the user's questions and emotional state, the emotion engine monitors the user's emotions and sends this information to the server.

[1012] Server: The server analyzes the emotional data and adjusts the tone and level of detail of the explanations and responses. For example, if the user is confused or stressed, the server will provide a more detailed and understandable explanation.

[1013] Terminal: Display the adjusted answer to the user. This process is repeated each time the user enters a new question.

[1014] Specific examples

[1015] For example, a user may upload a privacy policy document and request more information about "data retention period."

[1016] User: Uploads a privacy policy document via the web interface and fills in the form field with the topic "Data Retention Period".

[1017] Terminal: Sends document and topic information to the server. The user's emotional state is also monitored through the emotion engine.

[1018] Server: Analyzes the privacy policy document, extracts key sentences, and generates a summary. For example, selects the sentence "Data will be stored for a minimum of five years" as a key sentence. Re-analyzes the section on "Data storage period" and generates a detailed explanation, for example, "Data will be stored for a minimum of five years and then deleted." The summary and customized explanation are stored in a database and provided to the user in an optimized format based on the results of the sentiment engine.

[1019] Device: Display a summary and customized explanation to the user. For example, a dashboard displays a summary and detailed explanation.

[1020] Users: Check if it contains the information they want to know.

[1021] Server: Generates answers to questions as users enter follow-up questions in the conversational interface. Based on the analysis results of the emotion engine, the server adjusts the tone of the answers. For example, if the user is confused, it provides more detailed and polite explanations.

[1022] Terminal: Display the adjusted answer to the user.

[1023] Prompt Sentence Examples

[1024] An example of a prompt to input to a generative AI model could be, "Please summarize the privacy policy document you uploaded and provide detailed information about data retention periods."

[1025] This system allows users to quickly understand complex policy documents and efficiently access the information they need, while also minimizing user stress and confusion through the use of an emotion engine.

[1026] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1027] Step 1: User uploads document

[1028] User: Uploads a policy document through a web interface. For example, the user selects a PDF file and clicks the upload button.

[1029] Input: A policy document file selected by the user.

[1030] Output: The uploaded policy document file is transferred to the device.

[1031] Specific Action: The user selects a file using the file input field in the browser and presses the upload button.

[1032] Step 2: Send the policy document to the server

[1033] Terminal: Sends the uploaded policy document to the server. Displays the sending status with a progress bar.

[1034] Input: The policy document file uploaded by the user.

[1035] Output: The policy document file transferred to the server.

[1036] Specific behavior: The device sends the uploaded file as an HTTP request to a specific API endpoint, and displays a success message when the file is successfully sent to the server.

[1037] Step 3: Server launches NLP engine and analyzes

[1038] Server: Runs a natural language processing (NLP) engine to analyze the policy document, performing grammar analysis, keyword extraction, and weighting of each sentence.

[1039] Input: The policy document file transferred to the server.

[1040] Output: A list of extracted important sentences.

[1041] What it does: The server calls the analysis engine's API to analyze the text data of the document. The engine tokenizes the text, identifies grammatical patterns, and evaluates and extracts important phrases and sentences.

[1042] Step 4: Save important statements to the database

[1043] Server: Store the extracted important sentences in a database.

[1044] Input: A list of extracted important sentences.

[1045] Output: Key statements stored in the database.

[1046] Specific operation: The server inserts the extracted important sentences into the database as structured data.

[1047] Step 5: Generate a summary

[1048] Server: Generates a summary based on important sentences from the analysis results. The generated summary is formatted in a simple format.

[1049] Input: Key statements stored in the database.

[1050] Output: The generated summary.

[1051] Specific actions: The server sorts the important sentences, creates a summary in a concise paragraph format, and stores it in the database.

[1052] Step 6: Enter a customizable description

[1053] User: Enter a specific topic or keyword in the web interface.

[1054] Input: The topic or keyword entered by the user.

[1055] Output: Topics and keywords sent to the device.

[1056] Specific actions: The user enters keywords of interest into the search box on the interface and presses the search button.

[1057] Step 7: Send topic information to the server

[1058] Terminal: The input topic and keyword information is sent to the server. At the same time, the emotion engine monitors the user's emotional state.

[1059] Input: The topic or keyword entered by the user.

[1060] Output: The topics and keywords sent to the server.

[1061] Specific operation: The device sends an HTTP request containing the user's topic information and emotional state data to the server.

[1062] Step 8: Generate customization instructions

[1063] Server: Generates detailed explanations of relevant parts based on the input topic.

[1064] Input: The topics or keywords sent to the server.

[1065] Output: The generated customization description.

[1066] Specific operation: The server searches the database based on the topic or keyword, retrieves relevant information, and generates a detailed explanation. The generated explanation is stored in the database.

[1067] Step 9: Provide customization instructions

[1068] Terminal: Display the generated customization description to the user.

[1069] Input: The generated customization description.

[1070] Output: Customization instructions displayed to the user.

[1071] Specific operation: The terminal displays the customization explanation received from the server on the interface.

[1072] Step 10: Check for changes to the policy document

[1073] Server: Periodically check for changes to the policy document.

[1074] Input: The latest policy document.

[1075] Output: The changes detected.

[1076] Specific operation: The server checks the file's update date and time and hash value to determine whether there have been any changes.

[1077] Step 11: Update the summary and description

[1078] Server: If a change is detected, the new version is automatically analyzed and the summary and customization description are updated.

[1079] Input: The most recent policy document where a change was detected.

[1080] Output: Updated summary and commentary.

[1081] What happens: The server re-parses the changes, regenerates the summary and description, and stores them in the database.

[1082] Step 12: Notify users

[1083] Server: Sends notifications of updated information to users.

[1084] Input: Updated summary and commentary.

[1085] Output: The notification sent to the user.

[1086] What happens: The server displays a notification on the user's dashboard saying "Updates available."

[1087] Step 13: User enters question

[1088] User: Enters a question through a conversational interface.

[1089] Input: The question entered by the user.

[1090] Output: The question sent to the terminal.

[1091] Specific actions: The user enters a question into the question box on the interface and presses the submit button.

[1092] Step 14: Emotional state monitoring by the emotion engine

[1093] Terminal: Sends the user's question and emotional state to the server.

[1094] Input: User-entered question and emotional state data.

[1095] Output: Questions and emotional state data sent to the server.

[1096] Specific operation: The device analyzes the user's input using the emotion engine, generates emotional state data, and sends this data and the question to the server.

[1097] Step 15: Generate and refine explanations and answers

[1098] Server: Generates answers to questions and adjusts the answers based on sentiment data.

[1099] Input: User-entered question and emotional state data.

[1100] Output: The adjusted answer.

[1101] Specific operation: The server analyzes the question content and emotional state, generates a detailed explanation to make it easier to access, and adjusts the explanation based on the emotional data.

[1102] Step 16: Providing a tailored solution

[1103] Terminal: Display the adjusted answer to the user.

[1104] Input: The adjusted answer.

[1105] Output: The adjusted answer displayed in the interface.

[1106] Specific operation: The device displays the adjusted answer on the interface, allowing the user to confirm the answer.

[1107] The above are the processing steps of this system and their specific operations.

[1108] (Application example 2)

[1109] 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."

[1110] In autonomous vehicles, instruction manuals and operating instructions are complex, making it difficult for users to quickly and accurately obtain information. This often leads to confusion and stress. Furthermore, operating errors in such situations could lead to serious accidents. Therefore, there is a need for real-time, customized explanations and question responses that take users' emotions into consideration.

[1111] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1112] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing a detailed explanation; a means for detecting changes to the policy document and updating the summary and explanation; a means for the user to ask questions through an interactive interface and have the questions answered; a means for analyzing the document, extracting important sentences, and generating a summary; and a means for analyzing the user's emotions using an emotion engine and providing a detailed explanation according to the user's emotional state. This allows the user to quickly and accurately understand the instruction manual or operating manual for the autonomous vehicle and receive detailed explanations according to their emotional state, reducing confusion and stress and enabling safe and effective operation.

[1113] "Policy Document" refers to a document regarding various policies and regulations provided to users.

[1114] An "input means" is a device or interface for sending a policy document to a server.

[1115] "Natural language processing technology" is a computer technology for analyzing text data and understanding its meaning.

[1116] "Key Points" are information that is considered particularly important in a policy document.

[1117] An "extraction means" is a mechanism for selecting information from within a document based on certain criteria.

[1118] A "summarizing tool" is a method for concisely summarizing the extracted important points.

[1119] "User interests" are information that a user is particularly interested in or wants to know about.

[1120] "Detailed Description" refers to a detailed explanation of information provided based on the user's interest.

[1121] "Means for detecting changes" refers to the ability to automatically recognize changes to policy documents when they occur.

[1122] An "interactive interface" is a user interface that allows users to directly input questions and instructions into the system.

[1123] An "emotion engine" is a system for analyzing and understanding a user's emotional state.

[1124] "Emotional state" refers to the state of mind that a user feels in a particular situation.

[1125] The system of the present invention is designed to analyze, summarize, customize, and update policy documents, and to answer user questions. In this embodiment, the system is particularly applied to instruction manuals and operation manuals for autonomous vehicles.

[1126] System configuration and functions

[1127] 1. Document upload and text analysis

[1128] Users upload policy documents through input means such as smartphones or smart glasses.

[1129] The terminal transmits the uploaded policy document to the server.

[1130] The server uses natural language processing techniques to parse the policy document and extract key points, including grammar analysis, keyword extraction, and weighting each sentence.

[1131] The extracted important sentences are stored in a database.

[1132] 2. Summary Generation

[1133] The server generates a summary based on key sentences from the analysis results and provides it in a simple format that is easy for users to understand.

[1134] The generated summaries are stored in a database for quick access.

[1135] 3. Customizable commentary

[1136] Users enter topics or keywords of interest into an interactive interface.

[1137] The terminal sends this to the server.

[1138] Based on the topic entered, the server generates a detailed explanation of the relevant parts, including information on specific features and procedures.

[1139] 4. Regular updates

[1140] The server periodically checks for changes to the policy document and automatically parses the new version to update the summary and customization description.

[1141] The updated information will be notified to the user.

[1142] 5. User Interaction and Emotion Recognition

[1143] A user inputs a question through an interactive interface.

[1144] Example: "What are the conditions for airbag deployment?"

[1145] The device monitors the user's emotions through an emotion engine, which analyzes the user's emotional state and sends the results to the server.

[1146] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[1147] Specific examples

[1148] For example, if a user uploads a driverless vehicle instruction manual to their smart glasses and requests detailed information about "airbag deployment conditions," if the emotion engine recognizes "confusion," it will provide a polite response like this:

[1149] Example prompt sentence:

[1150] What are the conditions for airbag deployment?

[1151] Example answer:

[1152] Regarding the conditions for airbag deployment, the airbag system must first detect an emergency situation. It will deploy at a specific speed and angle of impact. Shall I explain this in more detail?

[1153] In this way, the emotion engine adjusts responses according to the user's emotional state, allowing the user to obtain more appropriate and reliable information, preventing operational errors and improving safety.

[1154] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1155] Step 1:

[1156] Users upload policy documents using their smartphones or smart glasses.

[1157] Input: Policy document

[1158] Output: The policy document is sent to the server.

[1159] Specific operation: The user selects a file through the device interface and clicks the upload button.

[1160] Step 2:

[1161] The terminal sends the uploaded policy document to the server.

[1162] Input: A user-uploaded policy document

[1163] Output: The policy document is transferred to the server.

[1164] Specific operation: By clicking the upload button, the device sends an HTTP request to the server.

[1165] Step 3:

[1166] The server uses natural language processing techniques to parse the policy document and extract key points.

[1167] Input: The policy document sent to the server

[1168] Output: Extracted key points

[1169] Specific operation: The server passes the policy document to an analysis tool, which performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[1170] Step 4:

[1171] The server generates a summary of the extracted key points.

[1172] Input: Key points extracted from the analysis

[1173] Output: Generated summary

[1174] What it does: Select the most important sentences and combine them to form a concise summary.

[1175] Step 5:

[1176] Users enter topics or keywords of interest into an interactive interface.

[1177] Input: Topics or keywords entered by the user

[1178] Output: Topics and keywords sent to the server

[1179] Specific actions: The user enters an item of interest into a text box on the interface and clicks the submit button.

[1180] Step 6:

[1181] The device sends the entered topics and keywords to the server.

[1182] Input: Topic or keyword

[1183] Output: Topics and keywords are sent to the server

[1184] Specific operation: By clicking the send button, the device sends an HTTP request to the server.

[1185] Step 7:

[1186] The server generates a detailed explanation of the relevant parts based on the topic entered.

[1187] Input: Topics or keywords sent to the server

[1188] Output: Detailed explanation generated

[1189] What happens: The server uses natural language processing techniques to re-parse the relevant section and generate a detailed explanation.

[1190] Step 8:

[1191] The server uses an emotion engine to analyze the user's emotions and provide detailed explanations according to their emotional state.

[1192] Input: User emotion data

[1193] Output: Detailed explanation according to emotional state

[1194] Specific operation: The emotion engine analyzes the user's emotional data and adjusts the content and tone of the commentary based on the results.

[1195] Step 9:

[1196] The server periodically checks for changes to the policy document and updates the summary and customization description.

[1197] Input: New policy document

[1198] Output: Updated summary and customization commentary

[1199] What it does: The server periodically reparses the document and updates the summary and detailed description if it detects any changes.

[1200] Step 10:

[1201] A user enters a question through an interactive interface and receives an answer to that question.

[1202] Input: User question

[1203] Output: The answer provided to the user

[1204] Specific operation: The user enters a question into a text box on the interface, clicks the submit button, and the server generates and displays the answer.

[1205] 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.

[1206] 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.

[1207] 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.

[1208] [Third embodiment]

[1209] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1210] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1211] 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).

[1212] 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.

[1213] 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.

[1214] 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).

[1215] 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.

[1216] 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.

[1217] 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.

[1218] 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.

[1219] 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.

[1220] 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."

[1221] The system of the present invention efficiently analyzes, summarizes, customizes, explains, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[1222] System configuration and functions

[1223] 1. Document upload and text analysis

[1224] User:

[1225] Users can upload policy documents through a web interface.

[1226] Device:

[1227] The policy document uploaded by the user is sent from the terminal to the server.

[1228] server:

[1229] The server runs a natural language processing (NLP) engine to parse the policy document.

[1230] The NLP engine performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[1231] Important statements are selected and stored in a database on the server.

[1232] 2. Summary Generation

[1233] server:

[1234] A summary is generated based on important sentences from the analysis results.

[1235] The generated summary is presented in a simple format that is easy for the user to understand.

[1236] The summaries are stored in a database for quick access at all times.

[1237] 3. Customizable commentary

[1238] User:

[1239] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[1240] Device:

[1241] The entered topics and keywords are sent from the terminal to the server.

[1242] server:

[1243] The server generates a detailed explanation of the relevant parts based on the topic entered.

[1244] The customized explanations are stored in a database and provided to the user.

[1245] 4. Regular updates

[1246] server:

[1247] The system periodically checks for changes to the policy document.

[1248] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[1249] The user is notified of any updated information, so they can always keep up to date with the latest information.

[1250] 5. User Interaction

[1251] User:

[1252] A conversational interface allows users to enter questions about the policy.

[1253] For example, you could ask, "How does this policy protect my personal information?"

[1254] Device:

[1255] The user's question is sent from the terminal to the server.

[1256] server:

[1257] The server re-parses the relevant sections based on the question and generates an answer.

[1258] The generated answer is sent to the terminal and provided to the user.

[1259] Specific examples

[1260] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[1261] User:

[1262] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[1263] Device:

[1264] The document and topic of interest information are sent to a server.

[1265] server:

[1266] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[1267] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[1268] The summaries and customized descriptions are stored in a database and sent to the user's terminal.

[1269] Device:

[1270] Display a summary and customizable description to the user.

[1271] User:

[1272] Review the information provided and familiarize yourself with the policy.

[1273] The system allows users to quickly understand complex policy documents and easily access important information and details of interest. It provides up-to-date information and immediate responses to user questions, significantly improving user convenience and comprehension.

[1274] The processing flow will be explained below.

[1275] Step 1:

[1276] Users upload policy documents through a web interface.

[1277] Step 2:

[1278] The device receives the uploaded policy document and sends this data to the server.

[1279] Step 3:

[1280] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[1281] Step 4:

[1282] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[1283] Step 5:

[1284] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[1285] Step 6:

[1286] The summarized results are stored in a database on the server for quick access later.

[1287] Step 7:

[1288] Users enter specific topics or keywords through a web interface, such as "data retention period."

[1289] Step 8:

[1290] The device sends the topics and keywords entered by the user to the server.

[1291] Step 9:

[1292] Based on the received topics and keywords, the server reparses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[1293] Step 10:

[1294] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[1295] Step 11:

[1296] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[1297] Step 12:

[1298] When there is an update, the server notifies the user, who can then access the latest information.

[1299] Step 13:

[1300] Users enter questions through a conversational interface, such as "How is my personal information protected?"

[1301] Step 14:

[1302] The device sends the user's question to the server.

[1303] Step 15:

[1304] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[1305] Step 16:

[1306] The generated answer is sent from the server to the device.

[1307] Step 17:

[1308] The device receives the answer from the server and displays it to the user.

[1309] The above is the specific operation performed at each processing step in this system, which enables users to quickly understand complex policy documents and access the information they need.

[1310] Example 1

[1311] 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."

[1312] Conventional policy document analysis systems have had many challenges in quickly and accurately extracting important information from documents and providing users with summaries and detailed explanations. Furthermore, they have limited functionality for responding to document changes in real time and providing immediate answers to user questions. This situation makes it difficult for users to understand the complex content of policies, requiring significant time and effort, and delays in updating information make it difficult to stay up to date with the latest information.

[1313] 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.

[1314] In this invention, the server includes a means for inputting a document, a means for analyzing the input document using natural language processing technology and extracting important information, a means for summarizing the extracted important information, a means for selecting specific information based on a user's request and providing a detailed explanation, a means for detecting changes in the document and updating the summary and explanation, and a means for the user to ask questions and receive answers through an interactive interface. This allows users to quickly understand complex policy documents and easily access important information and details of interest. Furthermore, the server is always provided with the latest information and can immediately respond to user questions, significantly improving user convenience and understanding.

[1315] A "document" is text data such as a policy or a regulation, which describes specific rules or information.

[1316] "Means for input" refers to an interface that allows a user to provide a document to the system, and includes file uploading and the like.

[1317] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and includes grammatical analysis, keyword extraction, semantic analysis, etc.

[1318] The "means for analyzing" is a method that provides a function for analyzing the contents of a document using natural language processing technology and extracting important information.

[1319] "Important information" refers to parts of a document that are particularly meaningful or valuable to the user.

[1320] "Means of extraction" refers to the technology used to select and extract important information from the analysis results.

[1321] "Methods of summarizing" are methods for concisely summarizing content based on extracted important information.

[1322] "Means for providing" refers to the method for displaying the processing results, summary, and detailed explanation to the user.

[1323] A "means for detecting changes" is a method that has the function of monitoring updates to a document and automatically recognizing changes.

[1324] "Means for updating" means techniques for updating the summary and detailed description with new content when changes are detected.

[1325] An "interactive interface" is an interface that allows a user to interact directly with a system, including chatbots for inputting questions and instructions.

[1326] "Answering means" refers to a technique for generating and providing an appropriate response to a question from a user.

[1327] The system of the present invention is a system for efficiently analyzing, summarizing, customizing explanations, updating policy documents, and answering user questions. An embodiment of this system is described in detail below.

[1328] System configuration and functions

[1329] Document upload and text analysis

[1330] User:

[1331] Users upload policy documents using a web interface, either by dragging and dropping files into the browser interface or by clicking the file chooser button.

[1332] Device:

[1333] The uploaded policy document is sent to the server via the terminal, using an HTTP POST request to send the file contents to the server in multipart format.

[1334] server:

[1335] The server temporarily stores the received document and starts a natural language processing (NLP) engine (e.g., spaCy). The server reads the document text and performs grammatical analysis, keyword extraction, and importance evaluation of each sentence. The analysis results are evaluated, and important sentences are identified and stored in a PostgreSQL database.

[1336] Generate a summary

[1337] server:

[1338] The server generates summaries based on key sentences from the analysis results of the NLP engine. It then uses algorithms such as TextRank and BERT to convert the summaries into a simple format that is easy for users to understand. The summaries are stored in a PostgreSQL database and can be quickly provided upon user request.

[1339] Customizable commentary

[1340] User:

[1341] A user types into the web interface to get information on a particular topic or keyword of interest, for example, typing "data retention period" into a text box.

[1342] Device:

[1343] The topics and keywords entered are sent to the server as an HTTP POST request.

[1344] server:

[1345] The server receives the request and generates a detailed explanation of the relevant part based on the input topic. The explanation is generated using a language model (e.g., OpenAI GPT-3). The customized explanation is stored in a database and returned to the user's device.

[1346] Regular updates

[1347] server:

[1348] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-parsed with the NLP engine, and the summary and customization descriptions are automatically updated. Updates are notified to users in real time through a notification system.

[1349] User Interaction

[1350] User:

[1351] Users use a conversational interface (e.g., a chatbot) to enter questions about the policy, such as "How does this policy protect my personal information?"

[1352] Device:

[1353] The user's question is sent directly to the server, and the input content is constructed as an HTTP request in natural language format and sent.

[1354] server:

[1355] The server receives the question, again using an NLP engine or language model to identify the relevant section, generates an answer to the question, and returns the result to the device, where the generated answer is provided to the user in real time.

[1356] Specific examples

[1357] For example, a specific flow will be described below when a user requests detailed information about the "data storage period."

[1358] User:

[1359] Upload your privacy policy document via the web interface, enter "Data Retention Period" as the topic of interest, and submit.

[1360] Device:

[1361] Submit the document and topic of interest information to the server as an HTTP POST request.

[1362] server:

[1363] The policy document is analyzed using spaCy, key sentences are extracted, and a summary is generated. The section regarding "data retention period" is then analyzed again, and a detailed explanation is generated using the OpenAI GPT-3 model. The generated summary and customized explanation are stored in a PostgreSQL database and sent to the device as an HTTP response.

[1364] Device:

[1365] Receives the response from the server and displays a summary and customized explanation to the user.

[1366] User:

[1367] Review the information provided to deepen your understanding of the policy. Depending on the topic, you may have more detailed questions or request information on a different topic.

[1368] Prompt Sentence Examples

[1369] "Please extract key statements from the uploaded policy document, generate a summary, and provide a detailed explanation of the 'data retention period'."

[1370] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1371] Step 1:

[1372] User:

[1373] Users upload policy documents through a web interface. Specifically, they can click the file chooser button in their browser to select a local file, or they can drag and drop a file into the interface. The input is the policy document (e.g., a text file or a PDF file). The output is the uploaded file information.

[1374] Step 2:

[1375] Device:

[1376] The uploaded policy document is sent from the terminal to the server. The file contents are sent to the server in multipart format using an HTTP POST request. The input is the local file path, and the output is a successful transfer to the server.

[1377] Step 3:

[1378] server:

[1379] The server temporarily stores the received document and starts a natural language processing (NLP) engine. The server reads the file contents, performs grammatical analysis, extracts keywords, and evaluates the importance of each sentence. The input is the uploaded file contents, and the output is the analysis results (a list of important sentences and their importance). Specifically, it uses the Python-based spaCy to perform grammatical analysis, extracts important keywords, and evaluates the importance of each sentence using the TextRank algorithm.

[1380] Step 4:

[1381] server:

[1382] A summary is generated based on important sentences from the analysis results. Algorithms such as TextRank and BERT are used to generate the summary. The generated summary is stored in a PostgreSQL database. The input is a list of important sentences, and the output is the generated summary. Specifically, it selects important sentences and converts them into a simple format to create a summary.

[1383] Step 5:

[1384] User:

[1385] A user enters information about a particular topic or keyword into a web interface. The input is typing the topic or keyword (e.g., "data retention period") into a text box. The output is sending that information to a server.

[1386] Step 6:

[1387] Device:

[1388] The topics and keywords entered by the user are sent to the server as an HTTP POST request. The input is the topic or keyword entered by the user, and the output is the completion of sending the request to the server.

[1389] Step 7:

[1390] server:

[1391] The server receives a request from the user and re-analyzes the relevant document portion based on the input topic. It uses a language model (e.g., OpenAI GPT-3) to generate a detailed explanation and stores the result in a database. The input is the topic, keywords, and the analysis result of the stored document, and the output is a customized detailed explanation. Specifically, it re-analyzes the relevant text and uses a generative AI model to create appropriate and detailed information for the user.

[1392] Step 8:

[1393] server:

[1394] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-analyzed by the NLP engine and the summary and customized commentary are updated. The input is the latest version of the document, which is periodically retrieved, and the output is the updated summary and commentary. Specifically, the system manages document versions, and if any changes occur, the latest data is retrieved and re-analyzed.

[1395] Step 9:

[1396] User:

[1397] Users use a conversational interface to enter questions about the policy. The input is a question entered through an interface such as a chatbot (e.g., "How does this policy protect my personal information?"), and the output is the entered question being sent to the server.

[1398] Step 10:

[1399] Device:

[1400] The user's question is sent from the terminal to the server. The input is the question entered by the user, and the output is the completion of sending the request to the server.

[1401] Step 11:

[1402] server:

[1403] The server receives the question and again uses an NLP engine or language model to identify relevant sections. It generates an answer to the question and returns the result to the device. The input is the question content and the analysis result of the relevant document parts, and the output is the generated answer. Specifically, it re-analyzes the relevant text and generates an appropriate answer to the user's question.

[1404] (Application example 1)

[1405] 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."

[1406] In physical stores, employee training and customer support require rapid understanding and response to policy documents, but this requires employees to quickly grasp the vast amount of content. However, with conventional methods, this process is cumbersome, and it is difficult to update information or respond quickly to specific questions. To solve these issues, it is necessary to develop a system that can summarize policy documents, provide customized explanations, and provide optimal answers to user questions.

[1407] 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.

[1408] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on a user's interest and providing a detailed explanation; a means for detecting changes in the policy document and updating the summary and explanation; a means for a user to ask questions and have the questions answered through an interactive interface; and a means for providing a summary and customized explanation of the policy document as an application to be installed on a mobile device or wearable device so that store employees can quickly understand the information and receive necessary training. This enables store employees to quickly understand the policy document, stay up to date with the latest information, and obtain accurate answers to specific questions.

[1409] A "policy document" is a document that deals with important policies and procedures such as laws, regulations, and guidelines.

[1410] An "input means" is an interface or device that allows a user to provide information to a system.

[1411] "Natural language processing technology" is a technology that enables computers to understand and process the language that humans use on a daily basis.

[1412] "Key points" are parts of a policy document that have particularly important content or meaning.

[1413] A "summary" is a short, concise summary of the original content.

[1414] "User interests" refer to specific topics or themes that a user is particularly interested in.

[1415] A "detailed explanation" is a specific and detailed explanation intended to provide a deeper understanding of the specific content.

[1416] The "means for detecting changes" is a function that automatically detects changes when the contents of a policy document are updated.

[1417] An "interactive interface" is an interface that allows a user to directly interact with a system and ask questions or give instructions.

[1418] A "mobile device" is a portable electronic device such as a smartphone or tablet.

[1419] A "wearable device" is an electronic device that is worn by the user.

[1420] The following describes a mode for realizing the present invention: The present invention provides a series of configurations and means as an application for mobile devices or wearable devices that enables employees of brick-and-mortar stores to quickly understand policy documents and receive the necessary training.

[1421] First, a user (employee) uses a mobile or wearable device such as a smartphone, tablet, or smart glasses. A dedicated application is installed on this device, and the user can input a policy document through its interface. The device then sends the input policy document to the server.

[1422] When the server receives the policy document, it uses natural language processing (NLP) technology to analyze the document. The NLP technology used includes specific software components such as SpaCy and Transformers (Hugging Face). The NLP engine on the server analyzes the grammar of each sentence and extracts keywords to evaluate and extract key points. The extracted key sentences are stored in a database.

[1423] The server then generates a summary based on the extracted key sentences. This summary is provided to the user in a simple, easy-to-understand format and can be viewed on their device. If the user requests further elaboration on a specific topic or keyword, the server uses a generative AI model, such as a BERT-based question-answering model, to generate a customized explanation based on the user's interests.

[1424] Additionally, the server periodically checks for changes to the policy document and automatically updates the summary and customized description according to any detected changes. This updated information is then posted back to the user's device.

[1425] When a user has a specific question, they send it to the server through a conversational interface. The server then analyzes the policy document again based on the question and generates the best answer. This process also leverages the generative AI model, generating prompts such as:

[1426] Input document:

[1427] "Our stores adhere to the following security policies: First, protecting customer information... (omitted)"

[1428] question:

[1429] "Please explain in more detail how you protect customer information."

[1430] This allows associates to quickly understand information, stay up-to-date on the latest policies, and get instant answers to specific questions, leading to efficient and effective operations in the physical store.

[1431] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1432] Step 1:

[1433] A user inputs a policy document using a mobile or wearable device such as a smartphone, tablet, or smart glasses. The policy document uploaded by the user through the application interface is then sent from the device to the server.

[1434] Input: Policy document (text file)

[1435] Output: Data sent to the server (policy document)

[1436] Step 2:

[1437] The server analyzes the received policy document using natural language processing (NLP) techniques, specifically using SpaCy and Transformers to perform grammatical analysis and keyword extraction, evaluate the importance of each sentence, and extract key points.

[1438] Input: Policy document

[1439] Data processing / data calculation: grammatical analysis, keyword extraction, importance evaluation

[1440] Output: Extracted important sentences (text data)

[1441] Step 3:

[1442] The server generates a summary based on the extracted key sentences, provides the summary in a simple format that is easy for users to understand, and stores it in a database.

[1443] Input: Extracted important sentences

[1444] Data processing / data calculation: Execution of summary generation algorithms

[1445] Output: Summary (text data)

[1446] Step 4:

[1447] When a user wants to get more information about a particular topic or keyword, they enter it through the application interface, and this information is sent from the device to the server.

[1448] Input: Topic or keyword

[1449] Output: Data sent to the server (topics and keywords)

[1450] Step 5:

[1451] The server generates detailed explanations of relevant parts based on topics and keywords provided by the user. It uses a generative AI model (e.g., a BERT-based question-answering model) to generate customized explanations and stores them in a database.

[1452] Input: Topics, keywords, policy documents

[1453] Data processing / data calculation: Generating detailed explanations (generating responses from generative AI models)

[1454] Output: Detailed explanation (text data)

[1455] Step 6:

[1456] When a user has a specific question, they send it to the server through the interactive interface, which then analyzes the question, generates an appropriate prompt, and parses the policy document again.

[1457] Input: User question

[1458] Data processing / data calculation: prompt generation, document re-analysis

[1459] Output: Prompt statement, part of analysis result

[1460] Step 7:

[1461] The server generates the best answer to the user's question and sends it to the device, where the user can check the answer on the application.

[1462] Input: prompt statement, analysis result

[1463] Data processing / data calculation: Answer generation (response generation for generative AI models)

[1464] Output: Answer (text data)

[1465] Step 8:

[1466] The server periodically checks for changes to the policy document, and if a change is detected, it parses the new version and automatically updates the summary and customization description. The updated information is notified to the user.

[1467] Input: Modified version of the policy document

[1468] Data processing / data calculations: difference checks, reanalysis, updating summaries and commentary

[1469] Output: Updated summary and commentary (text data)

[1470] Through these steps, the program helps store associates quickly understand policy documents and receive necessary training, while also providing immediate, targeted answers to specific questions.

[1471] 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.

[1472] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents combined with an emotion engine. An embodiment of the system is described in detail below.

[1473] System configuration and functions

[1474] 1. Document upload and text analysis

[1475] User:

[1476] Users can upload policy documents through a web interface.

[1477] Device:

[1478] The policy document uploaded by the user is sent from the terminal to the server.

[1479] server:

[1480] The server runs a natural language processing (NLP) engine to parse the policy document.

[1481] The NLP engine performs grammatical analysis and keyword extraction, applies a model that evaluates the importance of each sentence, and selects the most important sentences.

[1482] Important sentences are selected and stored in a database on the server.

[1483] 2. Summary Generation

[1484] server:

[1485] A summary is generated based on important sentences from the analysis results.

[1486] The generated summary is presented in a simple format that is easy for the user to understand.

[1487] The summaries are stored in a database for quick access at all times.

[1488] 3. Customizable commentary

[1489] User:

[1490] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[1491] Device:

[1492] The entered topics and keywords are sent from the terminal to the server.

[1493] server:

[1494] The server generates a detailed explanation of the relevant parts based on the topic entered.

[1495] The customized explanations are stored in a database and provided to the user.

[1496] 4. Regular updates

[1497] server:

[1498] The system periodically checks for changes to the policy document.

[1499] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[1500] The user is notified of any updated information, so they can always keep up to date with the latest information.

[1501] 5. User Interaction and Emotion Recognition

[1502] User:

[1503] A user can enter a question through an interactive interface.

[1504] For example, you could ask, "How does this policy protect my personal information?"

[1505] Device:

[1506] Based on the user's questions and usage, the device monitors the user's emotions through an emotion engine.

[1507] The emotion engine analyzes the user's emotional state and sends the results to the server.

[1508] server:

[1509] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[1510] Specific examples

[1511] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[1512] User:

[1513] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[1514] Device:

[1515] The document and topic of interest information are sent to the server, and the user's emotional state is also monitored through the emotion engine.

[1516] server:

[1517] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[1518] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[1519] The summaries and customized commentary are stored in a database and are presented to the user in an optimized format based on the results of the sentiment engine.

[1520] Device:

[1521] A summary and customized explanation are displayed to the user. When the user enters a question in the interactive interface, the server generates an answer that takes into account the user's emotional state.

[1522] server:

[1523] Answers to questions are generated and adjusted based on the user's emotional state.

[1524] Device:

[1525] The adjusted answer is displayed to the user.

[1526] This system not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state, further improving user convenience and understanding and reducing stress.

[1527] The processing flow will be explained below.

[1528] Step 1:

[1529] Users upload policy documents through a web interface.

[1530] Step 2:

[1531] The device receives the uploaded policy document and sends this data to the server.

[1532] Step 3:

[1533] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[1534] Step 4:

[1535] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[1536] Step 5:

[1537] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[1538] Step 6:

[1539] The summarized results are stored in a database on the server for quick access later.

[1540] Step 7:

[1541] Users enter specific topics or keywords through a web interface, such as "data retention period."

[1542] Step 8:

[1543] The device sends the topics and keywords entered by the user to the server.

[1544] Step 9:

[1545] Based on the received topics and keywords, the server re-parses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[1546] Step 10:

[1547] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[1548] Step 11:

[1549] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[1550] Step 12:

[1551] When there is an update, the server notifies the user, who can then access the latest information.

[1552] Step 13:

[1553] Users can enter questions through a conversational interface, such as "How is my personal information protected?"

[1554] Step 14:

[1555] The device sends the user's question to the server.

[1556] Step 15:

[1557] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[1558] Step 16:

[1559] When generating an answer, the emotion engine acquires emotion data from the user's device, allowing it to understand the emotional state of the user when they asked the question.

[1560] Step 17:

[1561] The server adjusts the tone and level of detail of the answer based on the emotional data. For example, if the user is feeling stressed, it will provide a more detailed and understandable explanation.

[1562] Step 18:

[1563] The generated answer is sent from the server to the device.

[1564] Step 19:

[1565] The device receives the answer from the server and displays it to the user.

[1566] Step 20:

[1567] Users can review the displayed answers and gain a deeper understanding of the policy content.

[1568] The above are the specific operations performed at each processing step in this system. This not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state. This further improves user convenience and understanding, and reduces stress.

[1569] Example 2

[1570] 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."

[1571] Conventional policy document analysis systems not only take time to understand complex documents and extract important information, but also have difficulty reducing user stress because they do not take into account the user's emotional state. Furthermore, when the content of a policy document is changed, updates are not automatically reflected, making it difficult to access the latest information. These problems need to be solved.

[1572] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing detailed explanations; a means for detecting changes to the policy document and updating the summaries and explanations; a means for the user to ask questions through an interactive interface and answer those questions; and a means for monitoring the user's emotional state through an emotion engine and adjusting the tone and level of detail of explanations and answers. This not only allows the user to quickly understand important information and easily access the latest information, but also enables polite responses according to the user's emotional state, thereby reducing stress for the user.

[1573] A "policy document" is a document that describes the policies, rules, and procedures of a company or organization.

[1574] "Natural language processing technology" refers to technology that allows computers to analyze, understand, and generate human language.

[1575] "Key Points" refers to information or elements within a policy document that deserve special attention.

[1576] "Extraction" refers to the process of extracting necessary information from documents or data.

[1577] An "abstract" is a brief description of the overall content of a document.

[1578] "User interests" refers to information or topics that are of particular interest or concern to a user.

[1579] "Specific portion" refers to a specific section or paragraph within a policy document.

[1580] A "detailed description" is a detailed and thorough description of a particular piece of information or topic.

[1581] "Change detection" refers to the ability to automatically identify content changes in policy documents.

[1582] "Updating" refers to the act of modifying data or descriptions to reflect new information or changes.

[1583] An "interactive interface" refers to an interface that allows a user to interact with a system to obtain information.

[1584] An "emotion engine" refers to a technology or system that analyzes a user's emotional state from their input and behavior and evaluates that state.

[1585] "Tone and detail adjustment" refers to the ability to appropriately change the wording and detail of explanations and answers provided depending on the user's emotional state.

[1586] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[1587] System configuration and functions

[1588] 1. Document upload and text analysis

[1589] User: Uploads a policy document through the web interface, or uses the file selector to select the document and clicks the upload button.

[1590] Terminal: Once a document is uploaded, it is sent to the server. It also has a function to display the progress according to the user's operations.

[1591] Server: Upon receiving the policy document, it launches a natural language processing (NLP) engine. The NLP engine performs grammatical analysis and keyword extraction, evaluates the importance of each sentence, and extracts important sentences. This extraction process uses technologies such as text analysis engines and machine learning models. Once the important sentences are extracted, they are stored in a database.

[1592] 2. Summary Generation

[1593] Server: The NLP engine generates a summary based on the extracted key sentences. The summary is formatted in a simple, easy-to-understand format and stored in the server's database, allowing users to quickly access it later.

[1594] 3. Customizable commentary

[1595] User: Enters a specific topic or keyword in the web interface. For example, the user enters an interest such as "data retention period."

[1596] Terminal: The input topic and keyword information is sent to the server, while the emotion engine monitors the user's emotional state.

[1597] Server: Based on the topic entered, the server will generate a detailed explanation of the relevant part. For example, it will re-analyze the section about "Data Retention Period" and generate a detailed explanation. This customized explanation will be stored in the database and provided to the user.

[1598] 4. Regular updates

[1599] Server: The system periodically checks for changes to the policy document, for example by checking the file modification date and time on a daily basis. If a change is detected, it automatically analyzes the new version and updates the summary and customization description.

[1600] Server: The updated information is saved in the database and notified to the user, who will see a notification message on their dashboard saying "There is an update."

[1601] 5. User Interaction and Emotion Recognition

[1602] User: Enters a question through a conversational interface, for example, a specific question such as "How does this policy protect my personal information?"

[1603] Device: Based on the user's questions and emotional state, the emotion engine monitors the user's emotions and sends this information to the server.

[1604] Server: The server analyzes the emotional data and adjusts the tone and level of detail of the explanations and responses. For example, if the user is confused or stressed, the server will provide a more detailed and understandable explanation.

[1605] Terminal: Display the adjusted answer to the user. This process is repeated each time the user enters a new question.

[1606] Specific examples

[1607] For example, a user may upload a privacy policy document and request more information about "data retention period."

[1608] User: Uploads a privacy policy document via the web interface and fills in the form field with the topic "Data Retention Period".

[1609] Terminal: Sends document and topic information to the server. The user's emotional state is also monitored through the emotion engine.

[1610] Server: Analyzes the privacy policy document, extracts key sentences, and generates a summary. For example, selects the sentence "Data will be stored for a minimum of five years" as a key sentence. Re-analyzes the section on "Data storage period" and generates a detailed explanation, for example, "Data will be stored for a minimum of five years and then deleted." The summary and customized explanation are stored in a database and provided to the user in an optimized format based on the results of the sentiment engine.

[1611] Device: Display a summary and customized explanation to the user. For example, a dashboard displays a summary and detailed explanation.

[1612] Users: Check if it contains the information they want to know.

[1613] Server: Generates answers to questions as users enter follow-up questions in the conversational interface. Based on the analysis results of the emotion engine, the server adjusts the tone of the answers. For example, if the user is confused, it provides more detailed and polite explanations.

[1614] Terminal: Display the adjusted answer to the user.

[1615] Prompt Sentence Examples

[1616] An example of a prompt to input to a generative AI model could be, "Please summarize the privacy policy document you uploaded and provide detailed information about data retention periods."

[1617] This system allows users to quickly understand complex policy documents and efficiently access the information they need, while also minimizing user stress and confusion through the use of an emotion engine.

[1618] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1619] Step 1: User uploads document

[1620] User: Uploads a policy document through a web interface. For example, the user selects a PDF file and clicks the upload button.

[1621] Input: A policy document file selected by the user.

[1622] Output: The uploaded policy document file is transferred to the device.

[1623] Specific Action: The user selects a file using the file input field in the browser and presses the upload button.

[1624] Step 2: Send the policy document to the server

[1625] Terminal: Sends the uploaded policy document to the server. Displays the sending status with a progress bar.

[1626] Input: The policy document file uploaded by the user.

[1627] Output: The policy document file transferred to the server.

[1628] Specific behavior: The device sends the uploaded file as an HTTP request to a specific API endpoint, and displays a success message when the file is successfully sent to the server.

[1629] Step 3: Server launches NLP engine and analyzes

[1630] Server: Runs a natural language processing (NLP) engine to analyze the policy document, performing grammar analysis, keyword extraction, and weighting of each sentence.

[1631] Input: The policy document file transferred to the server.

[1632] Output: A list of extracted important sentences.

[1633] What it does: The server calls the analysis engine's API to analyze the text data of the document. The engine tokenizes the text, identifies grammatical patterns, and evaluates and extracts important phrases and sentences.

[1634] Step 4: Save important statements to the database

[1635] Server: Store the extracted important sentences in a database.

[1636] Input: A list of extracted important sentences.

[1637] Output: Key statements stored in the database.

[1638] Specific operation: The server inserts the extracted important sentences into the database as structured data.

[1639] Step 5: Generate a summary

[1640] Server: Generates a summary based on important sentences from the analysis results. The generated summary is formatted in a simple format.

[1641] Input: Key statements stored in the database.

[1642] Output: The generated summary.

[1643] Specific actions: The server sorts the important sentences, creates a summary in a concise paragraph format, and stores it in the database.

[1644] Step 6: Enter a customizable description

[1645] User: Enter a specific topic or keyword in the web interface.

[1646] Input: The topic or keyword entered by the user.

[1647] Output: Topics and keywords sent to the device.

[1648] Specific actions: The user enters keywords of interest into the search box on the interface and presses the search button.

[1649] Step 7: Send topic information to the server

[1650] Terminal: The input topic and keyword information is sent to the server. At the same time, the emotion engine monitors the user's emotional state.

[1651] Input: The topic or keyword entered by the user.

[1652] Output: The topics and keywords sent to the server.

[1653] Specific operation: The device sends an HTTP request containing the user's topic information and emotional state data to the server.

[1654] Step 8: Generate customization instructions

[1655] Server: Generates detailed explanations of relevant parts based on the input topic.

[1656] Input: The topics or keywords sent to the server.

[1657] Output: The generated customization description.

[1658] Specific operation: The server searches the database based on the topic or keyword, retrieves relevant information, and generates a detailed explanation. The generated explanation is stored in the database.

[1659] Step 9: Provide customization instructions

[1660] Terminal: Display the generated customization description to the user.

[1661] Input: The generated customization description.

[1662] Output: Customization instructions displayed to the user.

[1663] Specific operation: The terminal displays the customization explanation received from the server on the interface.

[1664] Step 10: Check for changes to the policy document

[1665] Server: Periodically check for changes to the policy document.

[1666] Input: The latest policy document.

[1667] Output: The changes detected.

[1668] Specific operation: The server checks the file's update date and time and hash value to determine whether there have been any changes.

[1669] Step 11: Update the summary and description

[1670] Server: If a change is detected, the new version is automatically analyzed and the summary and customization description are updated.

[1671] Input: The most recent policy document where a change was detected.

[1672] Output: Updated summary and commentary.

[1673] What happens: The server re-parses the changes, regenerates the summary and description, and stores them in the database.

[1674] Step 12: Notify users

[1675] Server: Sends notifications of updated information to users.

[1676] Input: Updated summary and commentary.

[1677] Output: The notification sent to the user.

[1678] What happens: The server displays a notification on the user's dashboard saying "Updates available."

[1679] Step 13: User enters question

[1680] User: Enters a question through a conversational interface.

[1681] Input: The question entered by the user.

[1682] Output: The question sent to the terminal.

[1683] Specific actions: The user enters a question into the question box on the interface and presses the submit button.

[1684] Step 14: Emotional state monitoring by the emotion engine

[1685] Terminal: Sends the user's question and emotional state to the server.

[1686] Input: User-entered question and emotional state data.

[1687] Output: Questions and emotional state data sent to the server.

[1688] Specific operation: The device analyzes the user's input using the emotion engine, generates emotional state data, and sends this data and the question to the server.

[1689] Step 15: Generate and refine explanations and answers

[1690] Server: Generates answers to questions and adjusts the answers based on sentiment data.

[1691] Input: User-entered question and emotional state data.

[1692] Output: The adjusted answer.

[1693] Specific operation: The server analyzes the question content and emotional state, generates a detailed explanation to make it easier to access, and adjusts the explanation based on the emotional data.

[1694] Step 16: Providing a tailored solution

[1695] Terminal: Display the adjusted answer to the user.

[1696] Input: The adjusted answer.

[1697] Output: The adjusted answer displayed in the interface.

[1698] Specific operation: The device displays the adjusted answer on the interface, allowing the user to confirm the answer.

[1699] The above are the processing steps of this system and their specific operations.

[1700] (Application example 2)

[1701] 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."

[1702] In autonomous vehicles, instruction manuals and operating instructions are complex, making it difficult for users to quickly and accurately obtain information. This often leads to confusion and stress. Furthermore, operating errors in such situations could lead to serious accidents. Therefore, there is a need for real-time, customized explanations and question responses that take users' emotions into consideration.

[1703] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1704] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing a detailed explanation; a means for detecting changes to the policy document and updating the summary and explanation; a means for the user to ask questions through an interactive interface and have the questions answered; a means for analyzing the document, extracting important sentences, and generating a summary; and a means for analyzing the user's emotions using an emotion engine and providing a detailed explanation according to the user's emotional state. This allows the user to quickly and accurately understand the instruction manual or operating manual for the autonomous vehicle and receive detailed explanations according to their emotional state, reducing confusion and stress and enabling safe and effective operation.

[1705] "Policy Document" refers to a document regarding various policies and regulations provided to users.

[1706] An "input means" is a device or interface for sending a policy document to a server.

[1707] "Natural language processing technology" is a computer technology for analyzing text data and understanding its meaning.

[1708] "Key Points" are information that is considered particularly important in a policy document.

[1709] An "extraction means" is a mechanism for selecting information from within a document based on certain criteria.

[1710] A "summarizing tool" is a method for concisely summarizing the extracted important points.

[1711] "User interests" are information that a user is particularly interested in or wants to know about.

[1712] "Detailed Description" refers to a detailed explanation of information provided based on the user's interest.

[1713] "Means for detecting changes" refers to the ability to automatically recognize changes to policy documents when they occur.

[1714] An "interactive interface" is a user interface that allows users to directly input questions and instructions into the system.

[1715] An "emotion engine" is a system for analyzing and understanding a user's emotional state.

[1716] "Emotional state" refers to the state of mind that a user feels in a particular situation.

[1717] The system of the present invention is designed to analyze, summarize, customize, and update policy documents, and to answer user questions. In this embodiment, the system is particularly applied to instruction manuals and operation manuals for autonomous vehicles.

[1718] System configuration and functions

[1719] 1. Document upload and text analysis

[1720] Users upload policy documents through input means such as smartphones or smart glasses.

[1721] The terminal transmits the uploaded policy document to the server.

[1722] The server uses natural language processing techniques to parse the policy document and extract key points, including grammar analysis, keyword extraction, and weighting each sentence.

[1723] The extracted important sentences are stored in a database.

[1724] 2. Summary Generation

[1725] The server generates a summary based on key sentences from the analysis results and provides it in a simple format that is easy for users to understand.

[1726] The generated summaries are stored in a database for quick access.

[1727] 3. Customizable commentary

[1728] Users enter topics or keywords of interest into an interactive interface.

[1729] The terminal sends this to the server.

[1730] Based on the topic entered, the server generates a detailed explanation of the relevant parts, including information on specific features and procedures.

[1731] 4. Regular updates

[1732] The server periodically checks for changes to the policy document and automatically parses the new version to update the summary and customization description.

[1733] The updated information will be notified to the user.

[1734] 5. User Interaction and Emotion Recognition

[1735] A user inputs a question through an interactive interface.

[1736] Example: "What are the conditions for airbag deployment?"

[1737] The device monitors the user's emotions through an emotion engine, which analyzes the user's emotional state and sends the results to the server.

[1738] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[1739] Specific examples

[1740] For example, if a user uploads a driverless vehicle instruction manual to their smart glasses and requests detailed information about "airbag deployment conditions," if the emotion engine recognizes "confusion," it will provide a polite response like this:

[1741] Example prompt sentence:

[1742] What are the conditions for airbag deployment?

[1743] Example answer:

[1744] Regarding the conditions for airbag deployment, the airbag system must first detect an emergency situation. It will deploy at a specific speed and angle of impact. Shall I explain this in more detail?

[1745] In this way, the emotion engine adjusts responses according to the user's emotional state, allowing the user to obtain more appropriate and reliable information, preventing operational errors and improving safety.

[1746] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1747] Step 1:

[1748] Users upload policy documents using their smartphones or smart glasses.

[1749] Input: Policy document

[1750] Output: The policy document is sent to the server.

[1751] Specific operation: The user selects a file through the device interface and clicks the upload button.

[1752] Step 2:

[1753] The terminal sends the uploaded policy document to the server.

[1754] Input: A user-uploaded policy document

[1755] Output: The policy document is transferred to the server.

[1756] Specific operation: By clicking the upload button, the device sends an HTTP request to the server.

[1757] Step 3:

[1758] The server uses natural language processing techniques to parse the policy document and extract key points.

[1759] Input: The policy document sent to the server

[1760] Output: Extracted key points

[1761] Specific operation: The server passes the policy document to an analysis tool, which performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[1762] Step 4:

[1763] The server generates a summary of the extracted key points.

[1764] Input: Key points extracted from the analysis

[1765] Output: Generated summary

[1766] What it does: Select the most important sentences and combine them to form a concise summary.

[1767] Step 5:

[1768] Users enter topics or keywords of interest into an interactive interface.

[1769] Input: Topics or keywords entered by the user

[1770] Output: Topics and keywords sent to the server

[1771] Specific actions: The user enters an item of interest into a text box on the interface and clicks the submit button.

[1772] Step 6:

[1773] The device sends the entered topics and keywords to the server.

[1774] Input: Topic or keyword

[1775] Output: Topics and keywords are sent to the server

[1776] Specific operation: By clicking the send button, the device sends an HTTP request to the server.

[1777] Step 7:

[1778] The server generates a detailed explanation of the relevant parts based on the topic entered.

[1779] Input: Topics or keywords sent to the server

[1780] Output: Detailed explanation generated

[1781] What happens: The server uses natural language processing techniques to re-parse the relevant section and generate a detailed explanation.

[1782] Step 8:

[1783] The server uses an emotion engine to analyze the user's emotions and provide detailed explanations according to their emotional state.

[1784] Input: User emotion data

[1785] Output: Detailed explanation according to emotional state

[1786] Specific operation: The emotion engine analyzes the user's emotional data and adjusts the content and tone of the commentary based on the results.

[1787] Step 9:

[1788] The server periodically checks for changes to the policy document and updates the summary and customization description.

[1789] Input: New policy document

[1790] Output: Updated summary and customization commentary

[1791] What it does: The server periodically reparses the document and updates the summary and detailed description if it detects any changes.

[1792] Step 10:

[1793] A user enters a question through an interactive interface and receives an answer to that question.

[1794] Input: User question

[1795] Output: The answer provided to the user

[1796] Specific operation: The user enters a question into a text box on the interface, clicks the submit button, and the server generates and displays the answer.

[1797] 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.

[1798] 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.

[1799] 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.

[1800] [Fourth embodiment]

[1801] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1802] 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.

[1803] 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).

[1804] 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.

[1805] 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.

[1806] 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).

[1807] 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.

[1808] 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.

[1809] 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.

[1810] 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.

[1811] 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.

[1812] 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.

[1813] 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."

[1814] The system of the present invention efficiently analyzes, summarizes, customizes, explains, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[1815] System configuration and functions

[1816] 1. Document upload and text analysis

[1817] User:

[1818] Users can upload policy documents through a web interface.

[1819] Device:

[1820] The policy document uploaded by the user is sent from the terminal to the server.

[1821] server:

[1822] The server runs a natural language processing (NLP) engine to parse the policy document.

[1823] The NLP engine performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[1824] Important statements are selected and stored in a database on the server.

[1825] 2. Summary Generation

[1826] server:

[1827] A summary is generated based on important sentences from the analysis results.

[1828] The generated summary is presented in a simple format that is easy for the user to understand.

[1829] The summaries are stored in a database for quick access at all times.

[1830] 3. Customizable commentary

[1831] User:

[1832] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[1833] Device:

[1834] The entered topics and keywords are sent from the terminal to the server.

[1835] server:

[1836] The server generates a detailed explanation of the relevant parts based on the topic entered.

[1837] The customized explanations are stored in a database and provided to the user.

[1838] 4. Regular updates

[1839] server:

[1840] The system periodically checks for changes to the policy document.

[1841] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[1842] The user is notified of any updated information, so they can always keep up to date with the latest information.

[1843] 5. User Interaction

[1844] User:

[1845] A conversational interface allows users to enter questions about the policy.

[1846] For example, you could ask, "How does this policy protect my personal information?"

[1847] Device:

[1848] The user's question is sent from the terminal to the server.

[1849] server:

[1850] The server re-parses the relevant sections based on the question and generates an answer.

[1851] The generated answer is sent to the terminal and provided to the user.

[1852] Specific examples

[1853] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[1854] User:

[1855] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[1856] Device:

[1857] The document and topic of interest information are sent to a server.

[1858] server:

[1859] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[1860] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[1861] The summaries and customized descriptions are stored in a database and sent to the user's terminal.

[1862] Device:

[1863] Display a summary and customizable description to the user.

[1864] User:

[1865] Review the information provided and familiarize yourself with the policy.

[1866] The system allows users to quickly understand complex policy documents and easily access important information and details of interest. It provides up-to-date information and immediate responses to user questions, significantly improving user convenience and comprehension.

[1867] The processing flow will be explained below.

[1868] Step 1:

[1869] Users upload policy documents through a web interface.

[1870] Step 2:

[1871] The device receives the uploaded policy document and sends this data to the server.

[1872] Step 3:

[1873] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[1874] Step 4:

[1875] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[1876] Step 5:

[1877] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[1878] Step 6:

[1879] The summarized results are stored in a database on the server for quick access later.

[1880] Step 7:

[1881] Users enter specific topics or keywords through a web interface, such as "data retention period."

[1882] Step 8:

[1883] The device sends the topics and keywords entered by the user to the server.

[1884] Step 9:

[1885] Based on the received topics and keywords, the server reparses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[1886] Step 10:

[1887] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[1888] Step 11:

[1889] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[1890] Step 12:

[1891] When there is an update, the server notifies the user, who can then access the latest information.

[1892] Step 13:

[1893] Users enter questions through a conversational interface, such as "How is my personal information protected?"

[1894] Step 14:

[1895] The device sends the user's question to the server.

[1896] Step 15:

[1897] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[1898] Step 16:

[1899] The generated answer is sent from the server to the device.

[1900] Step 17:

[1901] The device receives the answer from the server and displays it to the user.

[1902] The above is the specific operation performed at each processing step in this system, which enables users to quickly understand complex policy documents and access the information they need.

[1903] Example 1

[1904] 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."

[1905] Conventional policy document analysis systems have had many challenges in quickly and accurately extracting important information from documents and providing users with summaries and detailed explanations. Furthermore, they have limited functionality for responding to document changes in real time and providing immediate answers to user questions. This situation makes it difficult for users to understand the complex content of policies, requiring significant time and effort, and delays in updating information make it difficult to stay up to date with the latest information.

[1906] 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.

[1907] In this invention, the server includes a means for inputting a document, a means for analyzing the input document using natural language processing technology and extracting important information, a means for summarizing the extracted important information, a means for selecting specific information based on a user's request and providing a detailed explanation, a means for detecting changes in the document and updating the summary and explanation, and a means for the user to ask questions and receive answers through an interactive interface. This allows users to quickly understand complex policy documents and easily access important information and details of interest. Furthermore, the server is always provided with the latest information and can immediately respond to user questions, significantly improving user convenience and understanding.

[1908] A "document" is text data such as a policy or a regulation, which describes specific rules or information.

[1909] "Means for input" refers to an interface that allows a user to provide a document to the system, and includes file uploading and the like.

[1910] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and includes grammatical analysis, keyword extraction, semantic analysis, etc.

[1911] The "means for analyzing" is a method that provides a function for analyzing the contents of a document using natural language processing technology and extracting important information.

[1912] "Important information" refers to parts of a document that are particularly meaningful or valuable to the user.

[1913] "Means of extraction" refers to the technology used to select and extract important information from the analysis results.

[1914] "Methods of summarizing" are methods for concisely summarizing content based on extracted important information.

[1915] "Means for providing" refers to the method for displaying the processing results, summary, and detailed explanation to the user.

[1916] A "means for detecting changes" is a method that has the function of monitoring updates to a document and automatically recognizing changes.

[1917] "Means for updating" means techniques for updating the summary and detailed description with new content when changes are detected.

[1918] An "interactive interface" is an interface that allows a user to interact directly with a system, including chatbots for inputting questions and instructions.

[1919] "Answering means" refers to a technique for generating and providing an appropriate response to a question from a user.

[1920] The system of the present invention is a system for efficiently analyzing, summarizing, customizing explanations, updating policy documents, and answering user questions. An embodiment of this system is described in detail below.

[1921] System configuration and functions

[1922] Document upload and text analysis

[1923] User:

[1924] Users upload policy documents using a web interface, either by dragging and dropping files into the browser interface or by clicking the file chooser button.

[1925] Device:

[1926] The uploaded policy document is sent to the server via the terminal, using an HTTP POST request to send the file contents to the server in multipart format.

[1927] server:

[1928] The server temporarily stores the received document and starts a natural language processing (NLP) engine (e.g., spaCy). The server reads the document text and performs grammatical analysis, keyword extraction, and importance evaluation of each sentence. The analysis results are evaluated, and important sentences are identified and stored in a PostgreSQL database.

[1929] Generate a summary

[1930] server:

[1931] The server generates summaries based on key sentences from the analysis results of the NLP engine. It then uses algorithms such as TextRank and BERT to convert the summaries into a simple format that is easy for users to understand. The summaries are stored in a PostgreSQL database and can be quickly provided upon user request.

[1932] Customizable commentary

[1933] User:

[1934] A user types into the web interface to get information on a particular topic or keyword of interest, for example, typing "data retention period" into a text box.

[1935] Device:

[1936] The topics and keywords entered are sent to the server as an HTTP POST request.

[1937] server:

[1938] The server receives the request and generates a detailed explanation of the relevant part based on the input topic. The explanation is generated using a language model (e.g., OpenAI GPT-3). The customized explanation is stored in a database and returned to the user's device.

[1939] Regular updates

[1940] server:

[1941] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-parsed with the NLP engine, and the summary and customization descriptions are automatically updated. Updates are notified to users in real time through a notification system.

[1942] User Interaction

[1943] User:

[1944] Users use a conversational interface (e.g., a chatbot) to enter questions about the policy, such as "How does this policy protect my personal information?"

[1945] Device:

[1946] The user's question is sent directly to the server, and the input content is constructed as an HTTP request in natural language format and sent.

[1947] server:

[1948] The server receives the question, again using an NLP engine or language model to identify the relevant section, generates an answer to the question, and returns the result to the device, where the generated answer is provided to the user in real time.

[1949] Specific examples

[1950] For example, a specific flow will be described below when a user requests detailed information about the "data storage period."

[1951] User:

[1952] Upload your privacy policy document via the web interface, enter "Data Retention Period" as the topic of interest, and submit.

[1953] Device:

[1954] Submit the document and topic of interest information to the server as an HTTP POST request.

[1955] server:

[1956] The policy document is analyzed using spaCy, key sentences are extracted, and a summary is generated. The section regarding "data retention period" is then analyzed again, and a detailed explanation is generated using the OpenAI GPT-3 model. The generated summary and customized explanation are stored in a PostgreSQL database and sent to the device as an HTTP response.

[1957] Device:

[1958] Receives the response from the server and displays a summary and customized explanation to the user.

[1959] User:

[1960] Review the information provided to deepen your understanding of the policy. Depending on the topic, you may have more detailed questions or request information on a different topic.

[1961] Prompt Sentence Examples

[1962] "Please extract key statements from the uploaded policy document, generate a summary, and provide a detailed explanation of the 'data retention period'."

[1963] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1964] Step 1:

[1965] User:

[1966] Users upload policy documents through a web interface. Specifically, they can click the file chooser button in their browser to select a local file, or they can drag and drop a file into the interface. The input is the policy document (e.g., a text file or a PDF file). The output is the uploaded file information.

[1967] Step 2:

[1968] Device:

[1969] The uploaded policy document is sent from the terminal to the server. The file contents are sent to the server in multipart format using an HTTP POST request. The input is the local file path, and the output is a successful transfer to the server.

[1970] Step 3:

[1971] server:

[1972] The server temporarily stores the received document and starts a natural language processing (NLP) engine. The server reads the file contents, performs grammatical analysis, extracts keywords, and evaluates the importance of each sentence. The input is the uploaded file contents, and the output is the analysis results (a list of important sentences and their importance). Specifically, it uses the Python-based spaCy to perform grammatical analysis, extracts important keywords, and evaluates the importance of each sentence using the TextRank algorithm.

[1973] Step 4:

[1974] server:

[1975] A summary is generated based on important sentences from the analysis results. Algorithms such as TextRank and BERT are used to generate the summary. The generated summary is stored in a PostgreSQL database. The input is a list of important sentences, and the output is the generated summary. Specifically, it selects important sentences and converts them into a simple format to create a summary.

[1976] Step 5:

[1977] User:

[1978] A user enters information about a particular topic or keyword into a web interface. The input is typing the topic or keyword (e.g., "data retention period") into a text box. The output is sending that information to a server.

[1979] Step 6:

[1980] Device:

[1981] The topics and keywords entered by the user are sent to the server as an HTTP POST request. The input is the topic or keyword entered by the user, and the output is the completion of sending the request to the server.

[1982] Step 7:

[1983] server:

[1984] The server receives a request from the user and re-analyzes the relevant document portion based on the input topic. It uses a language model (e.g., OpenAI GPT-3) to generate a detailed explanation and stores the result in a database. The input is the topic, keywords, and the analysis result of the stored document, and the output is a customized detailed explanation. Specifically, it re-analyzes the relevant text and uses a generative AI model to create appropriate and detailed information for the user.

[1985] Step 8:

[1986] server:

[1987] The system uses a cron job or task scheduler to periodically check for changes to policy documents. When changes are detected, the new document version is automatically re-analyzed by the NLP engine and the summary and customized commentary are updated. The input is the latest version of the document, which is periodically retrieved, and the output is the updated summary and commentary. Specifically, the system manages document versions, and if any changes occur, the latest data is retrieved and re-analyzed.

[1988] Step 9:

[1989] User:

[1990] Users use a conversational interface to enter questions about the policy. The input is a question entered through an interface such as a chatbot (e.g., "How does this policy protect my personal information?"), and the output is the entered question being sent to the server.

[1991] Step 10:

[1992] Device:

[1993] The user's question is sent from the terminal to the server. The input is the question entered by the user, and the output is the completion of sending the request to the server.

[1994] Step 11:

[1995] server:

[1996] The server receives the question and again uses an NLP engine or language model to identify relevant sections. It generates an answer to the question and returns the result to the device. The input is the question content and the analysis result of the relevant document parts, and the output is the generated answer. Specifically, it re-analyzes the relevant text and generates an appropriate answer to the user's question.

[1997] (Application example 1)

[1998] 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."

[1999] In physical stores, employee training and customer support require rapid understanding and response to policy documents, but this requires employees to quickly grasp the vast amount of content. However, with conventional methods, this process is cumbersome, and it is difficult to update information or respond quickly to specific questions. To solve these issues, it is necessary to develop a system that can summarize policy documents, provide customized explanations, and provide optimal answers to user questions.

[2000] 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.

[2001] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on a user's interest and providing a detailed explanation; a means for detecting changes in the policy document and updating the summary and explanation; a means for a user to ask questions and have the questions answered through an interactive interface; and a means for providing a summary and customized explanation of the policy document as an application to be installed on a mobile device or wearable device so that store employees can quickly understand the information and receive necessary training. This enables store employees to quickly understand the policy document, stay up to date with the latest information, and obtain accurate answers to specific questions.

[2002] A "policy document" is a document that deals with important policies and procedures such as laws, regulations, and guidelines.

[2003] An "input means" is an interface or device that allows a user to provide information to a system.

[2004] "Natural language processing technology" is a technology that enables computers to understand and process the language that humans use on a daily basis.

[2005] "Key points" are parts of a policy document that have particularly important content or meaning.

[2006] A "summary" is a short, concise summary of the original content.

[2007] "User interests" refer to specific topics or themes that a user is particularly interested in.

[2008] A "detailed explanation" is a specific and detailed explanation intended to provide a deeper understanding of the specific content.

[2009] The "means for detecting changes" is a function that automatically detects changes when the contents of a policy document are updated.

[2010] An "interactive interface" is an interface that allows a user to directly interact with a system and ask questions or give instructions.

[2011] A "mobile device" is a portable electronic device such as a smartphone or tablet.

[2012] A "wearable device" is an electronic device that is worn by the user.

[2013] The following describes a mode for realizing the present invention: The present invention provides a series of configurations and means as an application for mobile devices or wearable devices that enables employees of brick-and-mortar stores to quickly understand policy documents and receive the necessary training.

[2014] First, a user (employee) uses a mobile or wearable device such as a smartphone, tablet, or smart glasses. A dedicated application is installed on this device, and the user can input a policy document through its interface. The device then sends the input policy document to the server.

[2015] When the server receives the policy document, it uses natural language processing (NLP) technology to analyze the document. The NLP technology used includes specific software components such as SpaCy and Transformers (Hugging Face). The NLP engine on the server analyzes the grammar of each sentence and extracts keywords to evaluate and extract key points. The extracted key sentences are stored in a database.

[2016] The server then generates a summary based on the extracted key sentences. This summary is provided to the user in a simple, easy-to-understand format and can be viewed on their device. If the user requests further elaboration on a specific topic or keyword, the server uses a generative AI model, such as a BERT-based question-answering model, to generate a customized explanation based on the user's interests.

[2017] Additionally, the server periodically checks for changes to the policy document and automatically updates the summary and customized description according to any detected changes. This updated information is then posted back to the user's device.

[2018] When a user has a specific question, they send it to the server through a conversational interface. The server then analyzes the policy document again based on the question and generates the best answer. This process also leverages the generative AI model, generating prompts such as:

[2019] Input document:

[2020] "Our stores adhere to the following security policies: First, protecting customer information... (omitted)"

[2021] question:

[2022] "Please explain in more detail how you protect customer information."

[2023] This allows associates to quickly understand information, stay up-to-date on the latest policies, and get instant answers to specific questions, leading to efficient and effective operations in the physical store.

[2024] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[2025] Step 1:

[2026] A user inputs a policy document using a mobile or wearable device such as a smartphone, tablet, or smart glasses. The policy document uploaded by the user through the application interface is then sent from the device to the server.

[2027] Input: Policy document (text file)

[2028] Output: Data sent to the server (policy document)

[2029] Step 2:

[2030] The server analyzes the received policy document using natural language processing (NLP) techniques, specifically using SpaCy and Transformers to perform grammatical analysis and keyword extraction, evaluate the importance of each sentence, and extract key points.

[2031] Input: Policy document

[2032] Data processing / data calculation: grammatical analysis, keyword extraction, importance evaluation

[2033] Output: Extracted important sentences (text data)

[2034] Step 3:

[2035] The server generates a summary based on the extracted key sentences, provides the summary in a simple format that is easy for users to understand, and stores it in a database.

[2036] Input: Extracted important sentences

[2037] Data processing / data calculation: Execution of summary generation algorithms

[2038] Output: Summary (text data)

[2039] Step 4:

[2040] When a user wants to get more information about a particular topic or keyword, they enter it through the application interface, and this information is sent from the device to the server.

[2041] Input: Topic or keyword

[2042] Output: Data sent to the server (topics and keywords)

[2043] Step 5:

[2044] The server generates detailed explanations of relevant parts based on topics and keywords provided by the user. It uses a generative AI model (e.g., a BERT-based question-answering model) to generate customized explanations and stores them in a database.

[2045] Input: Topics, keywords, policy documents

[2046] Data processing / data calculation: Generating detailed explanations (generating responses from generative AI models)

[2047] Output: Detailed explanation (text data)

[2048] Step 6:

[2049] When a user has a specific question, they send it to the server through the interactive interface, which then analyzes the question, generates an appropriate prompt, and parses the policy document again.

[2050] Input: User question

[2051] Data processing / data calculation: prompt generation, document re-analysis

[2052] Output: Prompt statement, part of analysis result

[2053] Step 7:

[2054] The server generates the best answer to the user's question and sends it to the device, where the user can check the answer on the application.

[2055] Input: prompt statement, analysis result

[2056] Data processing / data calculation: Answer generation (response generation for generative AI models)

[2057] Output: Answer (text data)

[2058] Step 8:

[2059] The server periodically checks for changes to the policy document, and if a change is detected, it parses the new version and automatically updates the summary and customization description. The updated information is notified to the user.

[2060] Input: Modified version of the policy document

[2061] Data processing / data calculations: difference checks, reanalysis, updating summaries and commentary

[2062] Output: Updated summary and commentary (text data)

[2063] Through these steps, the program helps store associates quickly understand policy documents and receive necessary training, while also providing immediate, targeted answers to specific questions.

[2064] 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.

[2065] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents combined with an emotion engine. An embodiment of the system is described in detail below.

[2066] System configuration and functions

[2067] 1. Document upload and text analysis

[2068] User:

[2069] Users can upload policy documents through a web interface.

[2070] Device:

[2071] The policy document uploaded by the user is sent from the terminal to the server.

[2072] server:

[2073] The server runs a natural language processing (NLP) engine to parse the policy document.

[2074] The NLP engine performs grammatical analysis and keyword extraction, applies a model that evaluates the importance of each sentence, and selects the most important sentences.

[2075] Important sentences are selected and stored in a database on the server.

[2076] 2. Summary Generation

[2077] server:

[2078] A summary is generated based on important sentences from the analysis results.

[2079] The generated summary is presented in a simple format that is easy for the user to understand.

[2080] The summaries are stored in a database for quick access at all times.

[2081] 3. Customizable commentary

[2082] User:

[2083] Users enter topics of interest in a web interface to obtain information about specific topics or keywords.

[2084] Device:

[2085] The entered topics and keywords are sent from the terminal to the server.

[2086] server:

[2087] The server generates a detailed explanation of the relevant parts based on the topic entered.

[2088] The customized explanations are stored in a database and provided to the user.

[2089] 4. Regular updates

[2090] server:

[2091] The system periodically checks for changes to the policy document.

[2092] When a change is detected, the new version is automatically parsed and the summary and customization description are updated.

[2093] The user is notified of any updated information, so they can always keep up to date with the latest information.

[2094] 5. User Interaction and Emotion Recognition

[2095] User:

[2096] A user can enter a question through an interactive interface.

[2097] For example, you could ask, "How does this policy protect my personal information?"

[2098] Device:

[2099] Based on the user's questions and usage, the device monitors the user's emotions through an emotion engine.

[2100] The emotion engine analyzes the user's emotional state and sends the results to the server.

[2101] server:

[2102] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[2103] Specific examples

[2104] For example, a user may request a summary of the privacy policy and would like more information regarding "data retention periods."

[2105] User:

[2106] Upload your privacy policy document via the web interface and enter "data retention period" as a topic of interest.

[2107] Device:

[2108] The document and topic of interest information are sent to the server, and the user's emotional state is also monitored through the emotion engine.

[2109] server:

[2110] Analyzes privacy policy documents, extracts key sentences, and generates summaries.

[2111] Re-analyze the section on "Data Retention Period" and generate a detailed explanation.

[2112] The summaries and customized commentary are stored in a database and are presented to the user in an optimized format based on the results of the sentiment engine.

[2113] Device:

[2114] A summary and customized explanation are displayed to the user. When the user enters a question in the interactive interface, the server generates an answer that takes into account the user's emotional state.

[2115] server:

[2116] Answers to questions are generated and adjusted based on the user's emotional state.

[2117] Device:

[2118] The adjusted answer is displayed to the user.

[2119] This system not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state, further improving user convenience and understanding and reducing stress.

[2120] The processing flow will be explained below.

[2121] Step 1:

[2122] Users upload policy documents through a web interface.

[2123] Step 2:

[2124] The device receives the uploaded policy document and sends this data to the server.

[2125] Step 3:

[2126] The server inputs the received policy document into a natural language processing (NLP) engine and begins parsing it.

[2127] Step 4:

[2128] The server's NLP engine parses the policy document to extract keywords, then applies a model that evaluates the importance of each sentence to select the most important sentences.

[2129] Step 5:

[2130] The server generates a summary based on the selected key sentences, and the summary is constructed in a concise and easy-to-understand format.

[2131] Step 6:

[2132] The summarized results are stored in a database on the server for quick access later.

[2133] Step 7:

[2134] Users enter specific topics or keywords through a web interface, such as "data retention period."

[2135] Step 8:

[2136] The device sends the topics and keywords entered by the user to the server.

[2137] Step 9:

[2138] Based on the received topics and keywords, the server re-parses the relevant sections in the policy document, extracts the necessary information, and generates a detailed explanation.

[2139] Step 10:

[2140] The generated customized explanations are stored in a database on the server, allowing users to access them at any time.

[2141] Step 11:

[2142] The server periodically checks for updates to the policy document, and if there are any changes, it reparses the new document and updates the summary and customization description.

[2143] Step 12:

[2144] When there is an update, the server notifies the user, who can then access the latest information.

[2145] Step 13:

[2146] Users can enter questions through a conversational interface, such as "How is my personal information protected?"

[2147] Step 14:

[2148] The device sends the user's question to the server.

[2149] Step 15:

[2150] The server reparses the policy document based on the question, extracts relevant information, and generates an answer.

[2151] Step 16:

[2152] When generating an answer, the emotion engine acquires emotion data from the user's device, allowing it to understand the emotional state of the user when they asked the question.

[2153] Step 17:

[2154] The server adjusts the tone and level of detail of the answer based on the emotional data. For example, if the user is feeling stressed, it will provide a more detailed and understandable explanation.

[2155] Step 18:

[2156] The generated answer is sent from the server to the device.

[2157] Step 19:

[2158] The device receives the answer from the server and displays it to the user.

[2159] Step 20:

[2160] Users can review the displayed answers and gain a deeper understanding of the policy content.

[2161] The above are the specific operations performed at each processing step in this system. This not only enables users to quickly understand complex policy documents and access the information they need, but also enables them to respond appropriately according to their emotional state. This further improves user convenience and understanding, and reduces stress.

[2162] Example 2

[2163] 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."

[2164] Conventional policy document analysis systems not only take time to understand complex documents and extract important information, but also have difficulty reducing user stress because they do not take into account the user's emotional state. Furthermore, when the content of a policy document is changed, updates are not automatically reflected, making it difficult to access the latest information. These problems need to be solved.

[2165] The identification process by the identification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing detailed explanations; a means for detecting changes to the policy document and updating the summaries and explanations; a means for the user to ask questions through an interactive interface and answer those questions; and a means for monitoring the user's emotional state through an emotion engine and adjusting the tone and level of detail of explanations and answers. This not only allows the user to quickly understand important information and easily access the latest information, but also enables polite responses according to the user's emotional state, thereby reducing stress for the user.

[2166] A "policy document" is a document that describes the policies, rules, and procedures of a company or organization.

[2167] "Natural language processing technology" refers to technology that allows computers to analyze, understand, and generate human language.

[2168] "Key Points" refers to information or elements within a policy document that deserve special attention.

[2169] "Extraction" refers to the process of extracting necessary information from documents or data.

[2170] An "abstract" is a brief description of the overall content of a document.

[2171] "User interests" refers to information or topics that are of particular interest or concern to a user.

[2172] "Specific portion" refers to a specific section or paragraph within a policy document.

[2173] A "detailed description" is a detailed and thorough description of a particular piece of information or topic.

[2174] "Change detection" refers to the ability to automatically identify content changes in policy documents.

[2175] "Updating" refers to the act of modifying data or descriptions to reflect new information or changes.

[2176] An "interactive interface" refers to an interface that allows a user to interact with a system to obtain information.

[2177] An "emotion engine" refers to a technology or system that analyzes a user's emotional state from their input and behavior and evaluates that state.

[2178] "Tone and detail adjustment" refers to the ability to appropriately change the wording and detail of explanations and answers provided depending on the user's emotional state.

[2179] The system of the present invention efficiently analyzes, summarizes, customizes, updates, and answers user questions about policy documents, and an embodiment of the system is described in detail below.

[2180] System configuration and functions

[2181] 1. Document upload and text analysis

[2182] User: Uploads a policy document through the web interface, or uses the file selector to select the document and clicks the upload button.

[2183] Terminal: Once a document is uploaded, it is sent to the server. It also has a function to display the progress according to the user's operations.

[2184] Server: Upon receiving the policy document, it launches a natural language processing (NLP) engine. The NLP engine performs grammatical analysis and keyword extraction, evaluates the importance of each sentence, and extracts important sentences. This extraction process uses technologies such as text analysis engines and machine learning models. Once the important sentences are extracted, they are stored in a database.

[2185] 2. Summary Generation

[2186] Server: The NLP engine generates a summary based on the extracted key sentences. The summary is formatted in a simple, easy-to-understand format and stored in the server's database, allowing users to quickly access it later.

[2187] 3. Customizable commentary

[2188] User: Enters a specific topic or keyword in the web interface. For example, the user enters an interest such as "data retention period."

[2189] Terminal: The input topic and keyword information is sent to the server, while the emotion engine monitors the user's emotional state.

[2190] Server: Based on the topic entered, the server will generate a detailed explanation of the relevant part. For example, it will re-analyze the section about "Data Retention Period" and generate a detailed explanation. This customized explanation will be stored in the database and provided to the user.

[2191] 4. Regular updates

[2192] Server: The system periodically checks for changes to the policy document, for example by checking the file modification date and time on a daily basis. If a change is detected, it automatically analyzes the new version and updates the summary and customization description.

[2193] Server: The updated information is saved in the database and notified to the user, who will see a notification message on their dashboard saying "There is an update."

[2194] 5. User Interaction and Emotion Recognition

[2195] User: Enters a question through a conversational interface, for example, a specific question such as "How does this policy protect my personal information?"

[2196] Device: Based on the user's questions and emotional state, the emotion engine monitors the user's emotions and sends this information to the server.

[2197] Server: The server analyzes the emotional data and adjusts the tone and level of detail of the explanations and responses. For example, if the user is confused or stressed, the server will provide a more detailed and understandable explanation.

[2198] Terminal: Display the adjusted answer to the user. This process is repeated each time the user enters a new question.

[2199] Specific examples

[2200] For example, a user may upload a privacy policy document and request more information about "data retention period."

[2201] User: Uploads a privacy policy document via the web interface and fills in the form field with the topic "Data Retention Period".

[2202] Terminal: Sends document and topic information to the server. The user's emotional state is also monitored through the emotion engine.

[2203] Server: Analyzes the privacy policy document, extracts key sentences, and generates a summary. For example, selects the sentence "Data will be stored for a minimum of five years" as a key sentence. Re-analyzes the section on "Data storage period" and generates a detailed explanation, for example, "Data will be stored for a minimum of five years and then deleted." The summary and customized explanation are stored in a database and provided to the user in an optimized format based on the results of the sentiment engine.

[2204] Device: Display a summary and customized explanation to the user. For example, a dashboard displays a summary and detailed explanation.

[2205] Users: Check if it contains the information they want to know.

[2206] Server: Generates answers to questions as users enter follow-up questions in the conversational interface. Based on the analysis results of the emotion engine, the server adjusts the tone of the answers. For example, if the user is confused, it provides more detailed and polite explanations.

[2207] Terminal: Display the adjusted answer to the user.

[2208] Prompt Sentence Examples

[2209] An example of a prompt to input to a generative AI model could be, "Please summarize the privacy policy document you uploaded and provide detailed information about data retention periods."

[2210] This system allows users to quickly understand complex policy documents and efficiently access the information they need, while also minimizing user stress and confusion through the use of an emotion engine.

[2211] The flow of the identification process in the second embodiment will be described with reference to FIG.

[2212] Step 1: User uploads document

[2213] User: Uploads a policy document through a web interface. For example, the user selects a PDF file and clicks the upload button.

[2214] Input: A policy document file selected by the user.

[2215] Output: The uploaded policy document file is transferred to the device.

[2216] Specific Action: The user selects a file using the file input field in the browser and presses the upload button.

[2217] Step 2: Send the policy document to the server

[2218] Terminal: Sends the uploaded policy document to the server. Displays the sending status with a progress bar.

[2219] Input: The policy document file uploaded by the user.

[2220] Output: The policy document file transferred to the server.

[2221] Specific behavior: The device sends the uploaded file as an HTTP request to a specific API endpoint, and displays a success message when the file is successfully sent to the server.

[2222] Step 3: Server launches NLP engine and analyzes

[2223] Server: Runs a natural language processing (NLP) engine to analyze the policy document, performing grammar analysis, keyword extraction, and weighting of each sentence.

[2224] Input: The policy document file transferred to the server.

[2225] Output: A list of extracted important sentences.

[2226] What it does: The server calls the analysis engine's API to analyze the text data of the document. The engine tokenizes the text, identifies grammatical patterns, and evaluates and extracts important phrases and sentences.

[2227] Step 4: Save important statements to the database

[2228] Server: Store the extracted important sentences in a database.

[2229] Input: A list of extracted important sentences.

[2230] Output: Key statements stored in the database.

[2231] Specific operation: The server inserts the extracted important sentences into the database as structured data.

[2232] Step 5: Generate a summary

[2233] Server: Generates a summary based on important sentences from the analysis results. The generated summary is formatted in a simple format.

[2234] Input: Key statements stored in the database.

[2235] Output: The generated summary.

[2236] Specific actions: The server sorts the important sentences, creates a summary in a concise paragraph format, and stores it in the database.

[2237] Step 6: Enter a customizable description

[2238] User: Enter a specific topic or keyword in the web interface.

[2239] Input: The topic or keyword entered by the user.

[2240] Output: Topics and keywords sent to the device.

[2241] Specific actions: The user enters keywords of interest into the search box on the interface and presses the search button.

[2242] Step 7: Send topic information to the server

[2243] Terminal: The input topic and keyword information is sent to the server. At the same time, the emotion engine monitors the user's emotional state.

[2244] Input: The topic or keyword entered by the user.

[2245] Output: The topics and keywords sent to the server.

[2246] Specific operation: The device sends an HTTP request containing the user's topic information and emotional state data to the server.

[2247] Step 8: Generate customization instructions

[2248] Server: Generates detailed explanations of relevant parts based on the input topic.

[2249] Input: The topics or keywords sent to the server.

[2250] Output: The generated customization description.

[2251] Specific operation: The server searches the database based on the topic or keyword, retrieves relevant information, and generates a detailed explanation. The generated explanation is stored in the database.

[2252] Step 9: Provide customization instructions

[2253] Terminal: Display the generated customization description to the user.

[2254] Input: The generated customization description.

[2255] Output: Customization instructions displayed to the user.

[2256] Specific operation: The terminal displays the customization explanation received from the server on the interface.

[2257] Step 10: Check for changes to the policy document

[2258] Server: Periodically check for changes to the policy document.

[2259] Input: The latest policy document.

[2260] Output: The changes detected.

[2261] Specific operation: The server checks the file's update date and time and hash value to determine whether there have been any changes.

[2262] Step 11: Update the summary and description

[2263] Server: If a change is detected, the new version is automatically analyzed and the summary and customization description are updated.

[2264] Input: The most recent policy document where a change was detected.

[2265] Output: Updated summary and commentary.

[2266] What happens: The server re-parses the changes, regenerates the summary and description, and stores them in the database.

[2267] Step 12: Notify users

[2268] Server: Sends notifications of updated information to users.

[2269] Input: Updated summary and commentary.

[2270] Output: The notification sent to the user.

[2271] What happens: The server displays a notification on the user's dashboard saying "Updates available."

[2272] Step 13: User enters question

[2273] User: Enters a question through a conversational interface.

[2274] Input: The question entered by the user.

[2275] Output: The question sent to the terminal.

[2276] Specific actions: The user enters a question into the question box on the interface and presses the submit button.

[2277] Step 14: Emotional state monitoring by the emotion engine

[2278] Terminal: Sends the user's question and emotional state to the server.

[2279] Input: User-entered question and emotional state data.

[2280] Output: Questions and emotional state data sent to the server.

[2281] Specific operation: The device analyzes the user's input using the emotion engine, generates emotional state data, and sends this data and the question to the server.

[2282] Step 15: Generate and refine explanations and answers

[2283] Server: Generates answers to questions and adjusts the answers based on sentiment data.

[2284] Input: User-entered question and emotional state data.

[2285] Output: The adjusted answer.

[2286] Specific operation: The server analyzes the question content and emotional state, generates a detailed explanation to make it easier to access, and adjusts the explanation based on the emotional data.

[2287] Step 16: Providing a tailored solution

[2288] Terminal: Display the adjusted answer to the user.

[2289] Input: The adjusted answer.

[2290] Output: The adjusted answer displayed in the interface.

[2291] Specific operation: The device displays the adjusted answer on the interface, allowing the user to confirm the answer.

[2292] The above are the processing steps of this system and their specific operations.

[2293] (Application example 2)

[2294] 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."

[2295] In autonomous vehicles, instruction manuals and operating instructions are complex, making it difficult for users to quickly and accurately obtain information. This often leads to confusion and stress. Furthermore, operating errors in such situations could lead to serious accidents. Therefore, there is a need for real-time, customized explanations and question responses that take users' emotions into consideration.

[2296] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[2297] In this invention, the server includes: a means for inputting a policy document; a means for analyzing the input policy document using natural language processing technology and extracting important points; a means for summarizing the extracted important points; a means for selecting specific parts based on the user's interest and providing a detailed explanation; a means for detecting changes to the policy document and updating the summary and explanation; a means for the user to ask questions through an interactive interface and have the questions answered; a means for analyzing the document, extracting important sentences, and generating a summary; and a means for analyzing the user's emotions using an emotion engine and providing a detailed explanation according to the user's emotional state. This allows the user to quickly and accurately understand the instruction manual or operating manual for the autonomous vehicle and receive detailed explanations according to their emotional state, reducing confusion and stress and enabling safe and effective operation.

[2298] "Policy Document" refers to a document regarding various policies and regulations provided to users.

[2299] An "input means" is a device or interface for sending a policy document to a server.

[2300] "Natural language processing technology" is a computer technology for analyzing text data and understanding its meaning.

[2301] "Key Points" are information that is considered particularly important in a policy document.

[2302] An "extraction means" is a mechanism for selecting information from within a document based on certain criteria.

[2303] A "summarizing tool" is a method for concisely summarizing the extracted important points.

[2304] "User interests" are information that a user is particularly interested in or wants to know about.

[2305] "Detailed Description" refers to a detailed explanation of information provided based on the user's interest.

[2306] "Means for detecting changes" refers to the ability to automatically recognize changes to policy documents when they occur.

[2307] An "interactive interface" is a user interface that allows users to directly input questions and instructions into the system.

[2308] An "emotion engine" is a system for analyzing and understanding a user's emotional state.

[2309] "Emotional state" refers to the state of mind that a user feels in a particular situation.

[2310] The system of the present invention is designed to analyze, summarize, customize, and update policy documents, and to answer user questions. In this embodiment, the system is particularly applied to instruction manuals and operation manuals for autonomous vehicles.

[2311] System configuration and functions

[2312] 1. Document upload and text analysis

[2313] Users upload policy documents through input means such as smartphones or smart glasses.

[2314] The terminal transmits the uploaded policy document to the server.

[2315] The server uses natural language processing techniques to parse the policy document and extract key points, including grammar analysis, keyword extraction, and weighting each sentence.

[2316] The extracted important sentences are stored in a database.

[2317] 2. Summary Generation

[2318] The server generates a summary based on key sentences from the analysis results and provides it in a simple format that is easy for users to understand.

[2319] The generated summaries are stored in a database for quick access.

[2320] 3. Customizable commentary

[2321] Users enter topics or keywords of interest into an interactive interface.

[2322] The terminal sends this to the server.

[2323] Based on the topic entered, the server generates a detailed explanation of the relevant parts, including information on specific features and procedures.

[2324] 4. Regular updates

[2325] The server periodically checks for changes to the policy document and automatically parses the new version to update the summary and customization description.

[2326] The updated information will be notified to the user.

[2327] 5. User Interaction and Emotion Recognition

[2328] A user inputs a question through an interactive interface.

[2329] Example: "What are the conditions for airbag deployment?"

[2330] The device monitors the user's emotions through an emotion engine, which analyzes the user's emotional state and sends the results to the server.

[2331] The server adjusts the tone and level of detail of its explanations and responses based on the emotional data it receives from the emotion engine. For example, if it detects that the user is confused or stressed, it will provide a more detailed and understandable explanation.

[2332] Specific examples

[2333] For example, if a user uploads a driverless vehicle instruction manual to their smart glasses and requests detailed information about "airbag deployment conditions," if the emotion engine recognizes "confusion," it will provide a polite response like this:

[2334] Example prompt sentence:

[2335] What are the conditions for airbag deployment?

[2336] Example answer:

[2337] Regarding the conditions for airbag deployment, the airbag system must first detect an emergency situation. It will deploy at a specific speed and angle of impact. Shall I explain this in more detail?

[2338] In this way, the emotion engine adjusts responses according to the user's emotional state, allowing the user to obtain more appropriate and reliable information, preventing operational errors and improving safety.

[2339] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2340] Step 1:

[2341] Users upload policy documents using their smartphones or smart glasses.

[2342] Input: Policy document

[2343] Output: The policy document is sent to the server.

[2344] Specific operation: The user selects a file through the device interface and clicks the upload button.

[2345] Step 2:

[2346] The terminal sends the uploaded policy document to the server.

[2347] Input: A user-uploaded policy document

[2348] Output: The policy document is transferred to the server.

[2349] Specific operation: By clicking the upload button, the device sends an HTTP request to the server.

[2350] Step 3:

[2351] The server uses natural language processing techniques to parse the policy document and extract key points.

[2352] Input: The policy document sent to the server

[2353] Output: Extracted key points

[2354] Specific operation: The server passes the policy document to an analysis tool, which performs grammatical analysis and keyword extraction, and evaluates the importance of each sentence.

[2355] Step 4:

[2356] The server generates a summary of the extracted key points.

[2357] Input: Key points extracted from the analysis

[2358] Output: Generated summary

[2359] What it does: Select the most important sentences and combine them to form a concise summary.

[2360] Step 5:

[2361] Users enter topics or keywords of interest into an interactive interface.

[2362] Input: Topics or keywords entered by the user

[2363] Output: Topics and keywords sent to the server

[2364] Specific actions: The user enters an item of interest into a text box on the interface and clicks the submit button.

[2365] Step 6:

[2366] The device sends the entered topics and keywords to the server.

[2367] Input: Topic or keyword

[2368] Output: Topics and keywords are sent to the server

[2369] Specific operation: By clicking the send button, the device sends an HTTP request to the server.

[2370] Step 7:

[2371] The server generates a detailed explanation of the relevant parts based on the topic entered.

[2372] Input: Topics or keywords sent to the server

[2373] Output: Detailed explanation generated

[2374] What happens: The server uses natural language processing techniques to re-parse the relevant section and generate a detailed explanation.

[2375] Step 8:

[2376] The server uses an emotion engine to analyze the user's emotions and provide detailed explanations according to their emotional state.

[2377] Input: User emotion data

[2378] Output: Detailed explanation according to emotional state

[2379] Specific operation: The emotion engine analyzes the user's emotional data and adjusts the content and tone of the commentary based on the results.

[2380] Step 9:

[2381] The server periodically checks for changes to the policy document and updates the summary and customization description.

[2382] Input: New policy document

[2383] Output: Updated summary and customization commentary

[2384] What it does: The server periodically reparses the document and updates the summary and detailed description if it detects any changes.

[2385] Step 10:

[2386] A user enters a question through an interactive interface and receives an answer to that question.

[2387] Input: User question

[2388] Output: The answer provided to the user

[2389] Specific operation: The user enters a question into a text box on the interface, clicks the submit button, and the server generates and displays the answer.

[2390] 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.

[2391] 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.

[2392] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[2393] 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.

[2394] FIG. 9 is a diagram illustrating 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 actions 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.

[2395] 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.

[2396] 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).

[2397] 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.

[2398] 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."

[2399] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[2400] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[2401] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[2402] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[2403] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[2404] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[2405] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[2406] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may...

Claims

1. A policy document is input means; A means for analyzing the input policy document using natural language processing technology and extracting important points; a means of summarizing the key points extracted; A means of selecting specific parts based on user interest and providing detailed descriptions; A means to detect changes to policy documents and update summaries and descriptions; A means for users to ask and answer questions through a conversational interface; A system including:

2. A means for analyzing a policy document by evaluating the importance of each sentence and extracting important sentences; A means for generating a summary based on the extracted important sentences; The system of claim 1 .

3. an interface means for inputting specific topics or keywords based on the user's interests; A means of generating detailed explanations of relevant parts based on input topics and keywords; The system of claim 1 .

Images