system

Abstract

Provide a system. 【Solution means】 Means for receiving a document, Means for generating a summary from the document, Means for extracting important keywords from the summary, Means for tagging the document based on the extracted keywords, Means for classifying the document, Means for displaying the tags, Means for searching for documents classified based on the searched keywords, A system including the above.

Description

Technical Field

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

Background Art

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

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When sorting documents and files, conventionally, it has been done based on human subjectivity, so variations in classification criteria and the labor required for searching have been problems. Also, in situations where a large number of documents need to be classified quickly and accurately, efficient means have been lacking. As a result, a great deal of time and labor have been required to find the necessary documents. In addition, the processes of tagging and summary generation have been inefficient, reducing the productivity of the entire document management system.

Means for Solving the Problems

[0005] The present invention solves the above problems with a system that includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from the summaries, means for tagging documents based on the extracted keywords, means for classifying documents, means for displaying tags, and means for searching for documents classified based on searched keywords. Specifically, a summary is generated from a document using natural language processing technology, and important keywords are extracted from the summary. Then, tags are automatically generated based on the extracted keywords and attached to the documents. As a result, documents are classified according to a consistent standard, and users can quickly search for the documents they need. Furthermore, the generated tags are displayed to the user and can be modified as needed, thereby realizing flexible and efficient document management.

[0006] "Means of receiving documents" refers to an interface that allows users to upload digital documents and files to a server, and a function that stores those documents on the server.

[0007] "Means for generating summaries from documents" refers to algorithms and systems that use natural language processing techniques to shorten and display the main content of an received document.

[0008] "Methods for extracting important keywords from summaries" refers to algorithms and filtering techniques used to select words and phrases that represent the main topics or themes of a document from the generated summary.

[0009] "Means of tagging documents" refers to systems and algorithms that automatically generate labels and tags related to documents based on extracted keywords and link them to the documents.

[0010] "Means of classifying documents" refers to systems and protocols for sorting documents into specific categories or folders based on assigned tags or keywords.

[0011] "Means of displaying tags" refers to interfaces and display technologies that visually present generated tags and keywords to users, allowing users to review and modify their content.

[0012] "Means of searching for documents classified based on searched keywords" refers to search engines and algorithms that quickly search for relevant documents in a database based on search keywords entered by the user and present the results to the user. [Brief explanation of the drawing]

[0013] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] This is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] This is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] This shows an emotion map where multiple emotions are mapped. [Figure 10] This shows an emotion map where multiple emotions are mapped. [Figure 11] This is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12]It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Embodiment 2 when combined with an emotion engine. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when combined with an emotion engine.

Mode for Carrying Out the Invention

[0014] Hereinafter, an example of an embodiment of the system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0015] First, the language used in the following description will be explained.

[0016] In the following embodiments, the labeled processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0017] In the following embodiments, the labeled RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0018] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0019] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0020] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0021] [First Embodiment]

[0022] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0023] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0024] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0025] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0026] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0027] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0028] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0029] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0030] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0031] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0032] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0033] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0034] This invention relates to a system for automatically sorting and efficiently searching documents. This invention enables consistent document management that is not influenced by human subjectivity, and significantly improves search efficiency.

[0035] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[0036] Specific processing of the system

[0037] 1. User uploads documents

[0038] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[0039] 2. Server-driven summary generation

[0040] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, it uses summarization algorithms such as TextRank and BERT to shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0041] 3. Keyword extraction by the server

[0042] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[0043] 4. Tag generation by the server

[0044] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[0045] 5. Displaying tags on the device

[0046] The device displays a list of generated tags so that the user can review them. The user can then review the displayed tags and make any necessary corrections.

[0047] 6. Server-based document classification

[0048] The server categorizes documents based on the identified tags. Specifically, documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information." This information is stored in a database and used for later searches.

[0049] 7. User search and browsing

[0050] The user performs a search using specific keywords. The terminal sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the terminal, allowing the user to quickly find and view the documents they need.

[0051] Specific example

[0052] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[0053] summary:

[0054] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0055] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[0056] Users can review and modify these tags. Ultimately, the server categorizes documents based on these tags and stores them in the database. When a user searches using keywords such as "new customer," the server can quickly find relevant documents and display the results.

[0057] This allows for efficient document searching and management, without relying on human subjectivity.

[0058] The following describes the processing flow.

[0059] Step 1:

[0060] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[0061] Step 2:

[0062] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[0063] Step 3:

[0064] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[0065] Step 4:

[0066] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[0067] Step 5:

[0068] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[0069] Step 6:

[0070] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[0071] Step 7:

[0072] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[0073] Step 8:

[0074] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[0075] Step 9:

[0076] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[0077] Step 10:

[0078] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[0079] Step 11:

[0080] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[0081] Step 12:

[0082] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[0083] Step 13:

[0084] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[0085] Step 14:

[0086] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[0087] Step 15:

[0088] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[0089] Step 16:

[0090] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[0091] (Example 1)

[0092] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0093] Currently, many companies and individuals face the challenge of managing vast amounts of documents. In this situation, document classification and searching are often done manually, which is time-consuming and labor-intensive. Furthermore, management based on human subjectivity can lead to problems with the consistency of classification and the accuracy of searches. Therefore, there is a growing need for systems that can automatically organize documents and efficiently search them.

[0094] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0095] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for displaying the tags on a terminal, means for classifying documents based on the tags, and means for searching for documents classified based on the searched keywords. This enables automatic organization of documents, consistent document management, and efficient searching.

[0096] "Means of receiving documents" refers to the interface and communication means for users to upload document files to the server.

[0097] A "means for generating summaries from documents" refers to a summarization algorithm that uses natural language processing technology to summarize the content of a document, extract only the important information, and condense it into short sentences.

[0098] "Methods for extracting important keywords from summaries" refer to algorithms and processes for selecting frequently occurring words and semantically important words from generated summaries.

[0099] "Means for tagging documents based on extracted keywords" refers to algorithms and systems that automatically generate relevant tags based on extracted keywords.

[0100] "Means of displaying tags on a device" refers to a display screen and interface that allows users to check and modify the generated tags.

[0101] "Means for classifying documents based on tags" refers to algorithms and database management systems that automatically classify documents into appropriate categories using established tags.

[0102] "Means for searching for documents classified based on searched keywords" refers to a system and process that searches a database for relevant documents based on keywords entered by the user and displays the results.

[0103] This invention relates to a system for automatically sorting and efficiently searching documents. To implement this system, the following specific hardware and software are used to perform each processing step.

[0104] First, users utilize a dedicated web interface or application (e.g., a web browser) on a device such as a personal computer or smartphone. Through this, users can upload document files they wish to sort to the server. Specifically, using a common web browser such as Google Chrome® or Mozilla Firefox, users select the document they want to upload from a file selection dialog and click the "Upload" button.

[0105] The server receives document files sent by users and generates summaries of those documents using natural language processing (NLP) techniques. This process applies summarization algorithms such as TextRank and BERT. The server uses these algorithms to extract and analyze the key content of the document and create a concise summary.

[0106] Next, the server extracts key keywords from the generated summary. This process utilizes algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. This selects frequently occurring or semantically important words from the summary as keywords.

[0107] The server generates tags related to the document based on the extracted keywords. For example, if a sales report is uploaded, the server generates a summary such as "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this, it extracts keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies." Then, it generates tags such as "fiscal year 2023," "sales increase," and "new customers" from these keywords and attaches them to the document.

[0108] The device displays the generated tag list to the user. Through this display, the user can review the tags and modify or add them as needed. The display interface is intuitive and user-friendly, designed for easy operation.

[0109] The server then categorizes the documents based on the identified tags. This process utilizes a tag classification algorithm and a database management system. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[0110] Ultimately, when a user searches for a document using specific keywords, the terminal sends the entered keywords to the server. The server searches the database for the relevant documents and displays the search results on the terminal. This allows the user to quickly find and view the documents they need.

[0111] An example of a specific prompt message is as follows:

[0112] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0113] Please extract the key keywords from the summary above.

[0114] Please generate tags appropriate for the document in question.

[0115] Based on this, please categorize the document as a "sales report".

[0116] As described above, this system enables automatic document sorting and efficient searching. This allows for consistent document management and significantly improves search efficiency.

[0117] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0118] Step 1:

[0119] User upload of documents

[0120] Users upload document files using a dedicated web interface or application via a device such as a personal computer or smartphone. The user launches a browser, selects the document file from the file selection dialog, and clicks the "Upload" button.

[0121] Input: Document file selected by the user

[0122] Output: Document file sent to the server

[0123] Step 2:

[0124] Receiving documents from the server

[0125] The server receives document files sent by users. The received files are temporarily stored for subsequent processing.

[0126] Input: Uploaded document file

[0127] Output: Saved document file

[0128] Step 3:

[0129] Server-based summary generation

[0130] The server analyzes stored document files and generates summaries using natural language processing techniques. Specific algorithms used include TextRank and BERT. The server analyzes the document content, extracts key information, and generates a concise summary.

[0131] Input: Saved document file

[0132] Data processing: Summarization using natural language processing techniques (TextRank and BERT)

[0133] Output: Generated summary

[0134] Step 4:

[0135] Keyword extraction by the server

[0136] The server extracts key keywords from the generated summary. This process uses tools such as TF-IDF, Word2Vec, and SpaCy. It calculates the frequency and importance of words included in the summary and selects keywords.

[0137] Input: Generated summary

[0138] Data processing: Keyword extraction using TF-IDF, Word2Vec, and SpaCy.

[0139] Output: Extracted keyword list

[0140] Step 5:

[0141] Server-based tag generation

[0142] The server generates relevant tags based on the extracted keywords. It analyzes the keywords and creates tags based on specific conditions.

[0143] Input: Extracted keyword list

[0144] Data processing: Keyword-based tag generation algorithm

[0145] Output: Generated tag list

[0146] Step 6:

[0147] Tag display by device

[0148] The terminal displays a list of tags sent from the server to the user. The user reviews the displayed tags and makes corrections or additions as needed.

[0149] Input: Tag list sent from the server

[0150] Output: Tag list displayed to the user

[0151] Step 7:

[0152] Server-based document classification

[0153] The server classifies documents into specific categories based on confirmed tags. It uses a tag classification algorithm to assign documents to the appropriate categories and saves them to the database.

[0154] Input: Confirmed tag list

[0155] Data processing: Tag-based document classification algorithm

[0156] Output: Category information of documents stored in the database

[0157] Step 8:

[0158] User search and browsing

[0159] The user enters a specific keyword into the search bar on their device to search for documents. The device sends the entered keyword to the server, which searches its database for the relevant documents. The search results are displayed on the device, allowing the user to view the necessary documents.

[0160] Input: Search keywords entered by the user

[0161] Output: List of documents displayed as search results

[0162] Specific operation: User enters keywords into the search bar → Terminal sends keywords to the server → Server searches for documents → Search results are displayed on the terminal.

[0163] (Application Example 1)

[0164] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0165] Traditional document management systems suffer from the drawback of requiring many manual and time-consuming operations for document uploading, summary generation, keyword extraction, tagging, classification, and searching. Furthermore, for use in field settings such as logistics centers, efficient and intuitive operation is required, necessitating a flexible system utilizing smart devices. Smart search methods, such as voice commands, are also in demand.

[0166] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0167] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents, means for displaying tags, means for searching for classified documents based on searched keywords, means for reading documents via a smart device and uploading them to the cloud, and means for displaying search results on the smart device's display. This enables the automation of the document management process and efficient and intuitive document management using smart devices.

[0168] "Means of receiving documents" refers to the functions or devices that allow users to upload documents.

[0169] "Means for generating summaries from documents" refers to functions or devices that shorten and concisely express the main content of long documents.

[0170] "Means for extracting important keywords from a summary" refers to a function or device that selects important words from a generated summary.

[0171] "Means of tagging documents based on extracted keywords" refers to functions or devices that automatically assign labels related to documents based on keywords.

[0172] "Means of classifying documents" refers to functions or devices that categorize documents into specific categories based on their content.

[0173] "Means of displaying tags" refers to functions or devices that visually present tags assigned to a document to the user.

[0174] "Means for searching for documents classified based on searched keywords" refers to functions or devices that find appropriate documents based on keywords entered by the user.

[0175] "Means of reading documents via smart devices and uploading them to the cloud" refers to functions or devices that use smart glasses or other devices to scan documents and send them to a cloud server via the internet.

[0176] "Means of displaying search results on a smart device's display" refers to functions or devices that visually present search results to the user on the screen of a device such as smart glasses.

[0177] To implement this invention, the following system configuration and specific processing are used. The embodiment combines a server, a smart device (e.g., smart glasses), and cloud storage.

[0178] The server forms the core of a system that automatically sorts and efficiently searches documents, and has the following functions:

[0179] 1. Means of receiving documents:

[0180] Users use the smart glasses' camera to photograph documents, and the captured document data is uploaded to the cloud via the internet. The application on the smart glasses provides an interface for easily capturing documents and sending them to the server.

[0181] 2. Means for generating summaries from documents:

[0182] The server processes the received document and generates a summary using natural language processing techniques (e.g., TextRank or BERT). This summarization process shortens and concisely displays the main points of the document. For example, if a logistics instruction sheet is uploaded, a summary such as "Inventory list for Q2 2023 increased by 10% year-on-year" might be generated.

[0183] 3. Methods for extracting key keywords from summaries:

[0184] Next, the server extracts key keywords from the generated summary. This involves using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy) to identify and extract important concepts within the document.

[0185] 4. A means of tagging documents based on extracted keywords:

[0186] The system automatically generates relevant tags from the extracted keywords and adds them to the document. For example, tags such as "FY2023," "Q2," and "Inventory Increase" will be generated.

[0187] 5. Means of classifying documents:

[0188] The server categorizes documents based on the identified tags. This assigns documents to categories such as "periodical reports" and "inventory management." This information is stored in a database and used for later searches.

[0189] 6. Means of displaying tags:

[0190] The tag list generated by the server is displayed on the smart glasses' screen. The user can view the displayed tags and make corrections as needed.

[0191] 7. Means for searching for documents categorized based on searched keywords:

[0192] Users can use the voice command function of their smart glasses to perform searches using specific keywords. The smart glasses' display shows tags and summaries of relevant documents, allowing users to quickly find the documents they are looking for.

[0193] Hardware and software to use

[0194] Hardware: Smart glasses (e.g., Google Glass®)

[0195] software:

[0196] Natural language processing models: spaCy, TextRank, BERT

[0197] Cloud storage: AWS® S3, Google Cloud Storage

[0198] Specific example

[0199] When logistics center staff receive a new delivery order, they use smart glasses to photograph the document and upload it. The cloud server summarizes the document and generates tags. When they say the voice command "Search for new delivery orders," the smart glasses screen displays a list of relevant documents.

[0200] Example of a prompt

[0201] "Shipping Instructions: Generate a summary inventory list for Q2 2023 and extract tags. Output: Inventory List Summary: Inventory volume increased by 10% year-on-year in Q2 2023. Related tags: 2023, Q2, Inventory volume, 10% increase"

[0202] In this way, the document management process can be automated and streamlined, supporting the operations of the logistics center. This improves work efficiency and allows users to quickly find the documents they need.

[0203] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0204] Step 1:

[0205] The user uses smart glasses to photograph a document. This document data becomes the input. The user uses the smart glasses application to easily capture the document, and the captured document is uploaded to the cloud via the internet. The output of this process is the document data stored in cloud storage.

[0206] Step 2:

[0207] The server retrieves document data stored in cloud storage. This document data becomes the input, and the server uses natural language processing techniques (e.g., TextRank, BERT) to generate a summary of the document. The generated summary becomes the output. Specifically, it extracts the main content from the document and summarizes it in a shortened form.

[0208] Step 3:

[0209] The server extracts key keywords from the generated summary. This summary data becomes the input, and important concepts within the document are identified using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy). The extracted keywords become the output. Specifically, the frequency and context of the keywords are analyzed, and their importance is evaluated.

[0210] Step 4:

[0211] The server assigns tags to documents based on extracted keywords. This keyword data serves as input, and appropriate tags are automatically generated. The generated tag list is the output. Specifically, it searches the database for tags highly relevant to the keywords and assigns them to the documents.

[0212] Step 5:

[0213] The server categorizes documents. This tag list serves as input, and the server assigns documents to categories such as "periodical reports" and "inventory management." The category information of the classified documents is output. Specifically, it maps documents to the appropriate category based on their tags.

[0214] Step 6:

[0215] The server sends the generated tag list to the terminal, which displays it on the smart glasses' screen. This tag list serves as input, and the tags are displayed so that the user can see them. The displayed tag list becomes the output. Specifically, the tags are displayed on the smart glasses' HUD (Heads-Up Display).

[0216] Step 7:

[0217] The user performs a search using a specific keyword via voice command. This keyword becomes the input, and the server searches the database for relevant documents. The searched document list is the output. Specifically, the system quickly finds documents that match or are related to the search keyword from the database and displays the results on the smart glasses' display.

[0218] Through the above processes, users can efficiently manage documents using smart devices. Furthermore, the search function allows them to quickly find the documents they need.

[0219] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0220] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it is an invention that combines a sentiment engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results.

[0221] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[0222] Specific processing of the system

[0223] 1. User uploads documents

[0224] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[0225] 2. Server-driven summary generation

[0226] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, summarization algorithms such as TextRank and BERT are used to generate summaries that shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0227] 3. Keyword extraction by the server

[0228] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[0229] 4. Tag generation by the server

[0230] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[0231] 5. Acquisition of sentiment data by the server

[0232] The server is equipped with an emotion engine that generates emotion data from user input and operation history. For example, the emotion engine analyzes the user's emotion data based on information such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency.

[0233] 6. Server-side adjustment of tags and sentiment data

[0234] The server uses sentiment data generated by the sentiment engine to adjust document tagging and classification. For example, if a user is determined to be stressed based on their recent search history, the server will prioritize displaying documents related to relaxation.

[0235] 7. Displaying tags on the device

[0236] The device displays a list of generated tags so that the user can review them. The user can review the displayed tags and modify them as needed.

[0237] 8. Server-based document classification

[0238] The server categorizes documents based on established tags and sentiment data. Specifically, documents are assigned to categories such as "Sales Reports," "Annual," and "Customer Information." This information is stored in a database and used for later searches.

[0239] 9. User search and browsing

[0240] The user performs a search using specific keywords. The device sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the device, allowing the user to quickly find and view the documents they need. During this process, an emotion engine customizes the search results based on the user's emotions, prioritizing documents that meet the user's needs.

[0241] Specific example

[0242] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[0243] summary:

[0244] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0245] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[0246] The server is equipped with an emotion engine that generates emotion data from the user's operation history and search history. Based on this emotion data, tagging and document classification are adjusted.

[0247] Ultimately, when a user performs a search using keywords such as "new customer," the server can quickly search for relevant documents, taking sentiment data into consideration, and display the results. In this way, it becomes possible to efficiently search and manage documents without relying on human subjectivity.

[0248] The following describes the processing flow.

[0249] Step 1:

[0250] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[0251] Step 2:

[0252] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[0253] Step 3:

[0254] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[0255] Step 4:

[0256] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[0257] Step 5:

[0258] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[0259] Step 6:

[0260] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[0261] Step 7:

[0262] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[0263] Step 8:

[0264] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[0265] Step 9:

[0266] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[0267] Step 10:

[0268] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[0269] Step 11:

[0270] The server uses an emotion engine to generate emotion data from the user's operation history and search history. The emotion engine determines the user's emotional state and stores that data.

[0271] Step 12:

[0272] The server adjusts tagging and document classification based on emotional data. For example, if a user is feeling stressed, it will prioritize displaying relaxation-related documents.

[0273] Step 13:

[0274] Based on the tags determined by the server, the documents are classified into appropriate categories. Based on the tag information, the documents are assigned to categories such as "Business Report" and "Customer Information".

[0275] Step 14:

[0276] The server saves the classification results in the database. The classified document information is stored in the database to enable rapid search.

[0277] Step 15:

[0278] The user enters keywords in the search bar and conducts a search. The entered keywords are sent to the server through the terminal.

[0279] Step 16:

[0280] The server searches the database for the corresponding documents. The search engine matches the documents based on the keywords and extracts the relevant documents.

[0281] Step 17:

[0282] The server sends the search results to the terminal. The document list obtained as the search result is displayed to the user.

[0283] Step 18:

[0284] The user selects the required document and checks the detailed content. The user can click on the corresponding document from the displayed document list and view its detailed content.

[0285] Step 19:

[0286] The sentiment engine customizes the search results based on the user's sentiment data. For example, if the sentiment engine determines that the user is feeling stressed, relaxation-related documents are preferentially displayed.

[0287] (Example 2)

[0288] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0289] Conventional document management systems statically tag and classify documents, failing to reflect user emotions and intentions. As a result, it is difficult for users to quickly and accurately search for the documents they are looking for. Furthermore, there is a lack of methods to customize document search results and improve the user experience by utilizing sentiment data. This invention aims to achieve more efficient and customized document management and retrieval by dynamically adjusting document tagging and classification using user sentiment data.

[0290] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0291] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for acquiring user sentiment data, means for adjusting the classification of tags and documents based on the acquired sentiment data, means for classifying documents, means for displaying tags, and means for searching for classified documents based on searched keywords. This enables dynamic tagging and document classification based on user sentiment data.

[0292] "Means of receiving documents" refers to a system that sends documents uploaded by users to a server and has the function of receiving those documents on the server.

[0293] "Means for generating summaries from documents" refers to a function in which a server analyzes the entire content of a document and automatically generates a concise summary that highlights the main points.

[0294] "Method for extracting important keywords from a summary" refers to a function where the server analyzes the generated summary and extracts important words and phrases that represent the content of the document.

[0295] "A means of tagging documents based on extracted keywords" refers to a system that analyzes keywords and automatically assigns highly relevant tags to documents based on those keywords.

[0296] "Means for acquiring user sentiment data" refers to a function that analyzes the user's operation history and search history to collect data for inferring the user's emotional state.

[0297] "Means for adjusting tags and document classifications based on acquired sentiment data" refers to a system that dynamically changes document tags and classifications using collected sentiment data, performing appropriate tagging and document classification according to the user's emotional state.

[0298] "Means of classifying documents" refers to the function of dividing documents into specific categories or folders based on tags assigned to them.

[0299] "Means for displaying tags" refers to a function that visually displays tags generated by the server on the user interface, allowing users to review and modify them.

[0300] "Means for searching for documents classified based on searched keywords" refers to a function in which the server searches the database for relevant documents based on the search keywords entered by the user and displays the results.

[0301] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it incorporates an emotion engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results. The specific implementation method of this system is described below.

[0302] Hardware and software used

[0303] The server is the entity that performs document data storage, analysis, tag generation, sentiment analysis, and search processing. Typical software includes libraries and frameworks for natural language processing (NLP) (e.g., TextRank, BERT, SpaCy), algorithms such as TF-IDF and Word2Vec for keyword extraction, and a sentiment analysis engine.

[0304] The terminal is a device used by the user to upload documents, view tags, and receive search results. A general web browser or a dedicated application is used as the interface.

[0305] The user is the person who uses the system and performs operations such as uploading documents, viewing and modifying tags, and searching.

[0306] Overview of the process

[0307] Uploading of documents

[0308] The user uploads the documents or files to be sorted to the server through the terminal. The files can be easily selected using a dedicated web interface or application and sent to the server by clicking the upload button.

[0309] Generation of summaries

[0310] The server analyzes the received document and generates a summary using natural language processing techniques such as TextRank and BERT. This summary is used to briefly grasp the main content of the document.

[0311] Extraction of keywords

[0312] The server extracts important keywords from the generated summary using algorithms such as TF-IDF, Word2Vec, and SpaCy. As a result, words and phrases representing the content of the document are extracted.

[0313] Tag generation

[0314] The server generates tags related to the document based on the extracted keywords. Using a tag generation algorithm, it converts the extracted keywords into tags and assigns them to the document.

[0315] Acquisition of emotional data

[0316] The server is equipped with an emotion engine that acquires emotion data based on the user's operation history and search history. The emotion engine analyzes the user's emotions from data such as what kind of searches the user performed and which documents they viewed.

[0317] Adjusting tags and document classification

[0318] The server adjusts document tagging and classification based on acquired emotional data. If the server determines that the user is stressed, it prioritizes displaying documents with tags and classifications related to relaxation.

[0319] Tag display

[0320] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify them as needed.

[0321] Document classification

[0322] The server categorizes documents based on established tags and sentiment data. Documents are categorized into categories such as "Sales Reports," "Annual," and "Customer Information," and stored in the database.

[0323] Search and browsing

[0324] The user enters specific keywords to perform a search. The server searches the database for relevant documents and sends the results to the user's terminal. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[0325] Specific example

[0326] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this summary, keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies" are extracted, and tags such as "fiscal year 2023," "sales increase," and "new customers" are generated.

[0327] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is seeking relaxation and will prioritize displaying related documents.

[0328] Examples of prompts for generative AI models

[0329] "After uploading the Sales Report.pdf file, perform text summarization, keyword extraction, and tag generation, then adjust the tags and document classification using sentiment data."

[0330] Thus, the system of the present invention can significantly improve the efficiency of document management and retrieval by utilizing user sentiment data.

[0331] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0332] Step 1:

[0333] User upload of documents

[0334] Users upload documents and files they want to categorize to the server using their device. Users open a dedicated web interface or application, click the file selection button, and select files such as "Sales Report.pdf". The selected files are sent from the device to the server by clicking the upload button.

[0335] Input: A document file uploaded by the user.

[0336] Output: Document file sent to the server.

[0337] Step 2:

[0338] Server-based summary generation

[0339] The server receives the uploaded document and determines the file format (PDF, Word, text, etc.). It extracts the document content using an appropriate parser and generates a document summary using natural language processing techniques (TextRank or BERT). The server analyzes the document content and creates a concise summary that includes the main points.

[0340] Input: Document file sent to the server.

[0341] Output: A summary of the document.

[0342] Specific example: The server analyzes "Sales Report.pdf" and generates a summary stating, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0343] Step 3:

[0344] Keyword extraction by the server

[0345] The server extracts key keywords from the generated summary. It uses algorithms such as TF-IDF, Word2Vec, and SpaCy to analyze the summary and select frequently occurring and important words.

[0346] Input: A summary of the document.

[0347] Output: A list of extracted important keywords.

[0348] Specific example: Extract keywords such as "FY2023," "First Quarter," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" from the summary.

[0349] Step 4:

[0350] Server-based tag generation

[0351] The server generates tags related to the document based on the extracted keywords. It converts the keywords into the appropriate tag format and assigns them to the document. A tag generation algorithm is used to automatically create relevant tags from the extracted keywords.

[0352] Input: A list of important keywords extracted.

[0353] Output: A list of tags assigned to the document.

[0354] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are generated.

[0355] Step 5:

[0356] Acquisition of emotional data by a server

[0357] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. It analyzes data such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency, in order to infer the user's emotional state.

[0358] Input: Data such as user activity history, search history, referenced documents, time of day, and frequency.

[0359] Output: Generated user sentiment data.

[0360] Specific example: If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is experiencing stress.

[0361] Step 6:

[0362] Server-based adjustment of tag and document classification

[0363] The server adjusts document tagging and classification based on the acquired sentiment data. It re-evaluates tags based on the sentiment data and makes changes or additions as needed.

[0364] Input: Generated user sentiment data, list of tags assigned to the document.

[0365] Output: Tag lists and categorized documents adjusted based on sentiment data.

[0366] Specific example: If a user is determined to be experiencing stress, documents with tags or categories related to relaxation will be prioritized for display.

[0367] Step 7:

[0368] Tag display by device

[0369] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify or add them as needed.

[0370] Input: A list of tags adjusted based on sentiment data.

[0371] Output: A list of tags displayed on the user's screen.

[0372] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are displayed on the user's screen.

[0373] Step 8:

[0374] Server-based document classification

[0375] The server categorizes documents based on established tags and sentiment data. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[0376] Input: Tag list adjusted based on sentiment data, document file.

[0377] Output: Category information for classified documents.

[0378] Specific example: "Sales Report.pdf" is categorized as "Sales Report," "Fiscal Year 2023," etc.

[0379] Step 9:

[0380] User search and browsing

[0381] The user enters specific keywords to perform a search. The terminal sends the entered keywords to the server, which searches its database for relevant documents. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[0382] Input: Search keywords entered by the user.

[0383] Output: Customized search results list.

[0384] Specific example: When a user searches using the keyword "new customer," relevant documents are displayed preferentially, and "Sales Report.pdf" is found among them.

[0385] (Application Example 2)

[0386] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0387] In today's information society, a vast amount of documents are stored electronically, making it difficult to efficiently search for necessary documents due to their sheer volume. Furthermore, conventional document management systems rely solely on simple keyword matching without considering the user's emotional state, making it difficult to quickly provide the information the user truly needs. Therefore, there is a need for a system that utilizes user emotional data to customize search results and provide a more personalized search experience.

[0388] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0389] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents using the tags, means for generating user sentiment data, means for adjusting document tags based on sentiment data, means for displaying the generated tags, and means for searching for documents classified based on searched keywords and adjusted tags. This enables automatic document classification and customization of search results based on user sentiment.

[0390] "Means of receiving documents" refers to a function that allows users to upload document data to a server via their device.

[0391] "Means for generating summaries from documents" refers to functions that use natural language processing technology to shorten and concisely display the main content of a document.

[0392] "Methods for extracting important keywords" refer to functions that extract frequently occurring and important information from document summaries.

[0393] "Means of tagging documents" refers to a function that automatically generates and assigns tags related to a document based on extracted keywords.

[0394] "A means of classifying documents using tags" refers to a function that sorts documents into specific categories based on the tags that have been generated.

[0395] "Means for generating user emotion data" refers to a function that analyzes a user's search history and operation history and generates data on their emotional state.

[0396] "Means for adjusting document tags based on sentiment data" refers to a function that dynamically changes the tags and classifications assigned to a document based on generated sentiment data.

[0397] "Means for displaying generated tags" refers to a function that displays a list of tags so that users can review and modify them.

[0398] "Means for searching for documents categorized based on searched keywords and adjusted tags" refers to a function that searches a database for appropriate documents based on keywords and adjusted tags entered by the user and displays the results.

[0399] The system that realizes this invention is designed to allow users to search for documents in an efficient and personalized manner. Specific embodiments are described below.

[0400] Hardware and software to be used

[0401] This system is recommended to use the following hardware and software.

[0402] Hardware:

[0403] CPU: Intel i7 or higher recommended

[0404] Memory: 16GB or more

[0405] Storage: 500GB SSD

[0406] User devices: PCs, smartphones, tablets, etc.

[0407] software:

[0408] OS: Windows 10 or Linux (registered trademark) Ubuntu 20.04

[0409] Libraries: SpaCy (en_core_web_sm model), TextBlob, Python 3.x

[0410] System configuration and operation

[0411] 1. Upload document

[0412] Users upload documents and files they want to categorize to the server via their device.

[0413] A dedicated web interface or smartphone app can be used.

[0414] 2. Summary generation

[0415] The server receives the uploaded document and generates a summary using natural language processing technology (such as SpaCy).

[0416] For example, summarization algorithms such as TextRank and BERT are used to shorten and concisely display the main content of a document.

[0417] 3. Keyword Extraction

[0418] The process of extracting key keywords from the summary uses TF-IDF, Word2Vec, and SpaCy.

[0419] For example, keywords such as "sales revenue" and "new customers" can be extracted from the content of uploaded documents.

[0420] 4. Tag generation

[0421] Based on the extracted keywords, generate tags relevant to the document.

[0422] This automatically adds tags such as "increased sales" and "new customers" to documents.

[0423] 5. Generation and adjustment of emotional data

[0424] The server generates sentiment data using a sentiment engine based on the user's operation history and search history.

[0425] For example, the document's tags are adjusted based on whether the user has recently viewed many negative reviews.

[0426] 6. Displaying tags and classifying documents

[0427] The device displays the generated tag list so that the user can review it and make corrections as needed.

[0428] The server categorizes documents into specific categories based on confirmed tags and sentiment data.

[0429] 7. Search and Browsing

[0430] When a user performs a search using a specific keyword, the device sends that keyword to the server.

[0431] The server searches the database for relevant documents and displays customized search results based on sentiment data.

[0432] Specific example

[0433] For example, if a user uploads a "Sales Report for the First Quarter of Fiscal Year 2023," the server receives the document, generates a summary, and displays a summary like the following:

[0434] summary:

[0435] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0436] Furthermore, keywords such as "FY2023," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" are extracted from the summary, and tags are generated based on these. After tag generation, if the user has recently seen many positive reviews, a "Positive" tag is added, adjusting the search results according to the user's sentiment.

[0437] Example of a prompt

[0438] Please write the following program to generate a user sentiment score based on reviews and adjust the document tags accordingly.

[0439] 1. Use Spasi to summarize the document.

[0440] 2. Extract keywords from the summary.

[0441] 3. Tags are generated based on the keywords extracted by Spasi.

[0442] 4. Calculate sentiment data from user reviews. (Examples: 'This product is amazing!', 'Not satisfied', 'Excellent quality')

[0443] 5. Adjust the tags generated based on the calculated sentiment data. (Add a "negative" tag if the sentiment is negative, and a "positive" tag if it's positive.)

[0444] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0445] Step 1:

[0446] The server receives documents uploaded by users via their terminals. The server receives the document data (PDF, DOCX, TXT, etc.) as input and stores it within the document management system. At the same time, the document's metadata (title, creation date, etc.) is also saved.

[0447] Step 2:

[0448] The server analyzes the received document data using natural language processing techniques (such as SpaCy) and generates a document summary. It analyzes the document text as input, extracts the main content, and generates a summary. It generates a summary text as output and passes it on to the next process. Specifically, this involves sentence segmentation, importance calculation, and selection of summary sentences.

[0449] Step 3:

[0450] The server extracts key keywords from the generated summary. Based on the summary text as input, it selects key keywords using algorithms such as TF-IDF and Word2Vec. It then generates a list of key keywords as output. Specifically, this involves calculating word frequency, applying inverse document frequency, and calculating weights.

[0451] Step 4:

[0452] The server automatically generates and assigns tags related to the document based on the extracted keywords. It uses a keyword list as input to select tags to be assigned to the document. It generates a tag list as output and associates it with the document. Specifically, this involves keyword mapping and the generation of category tags.

[0453] Step 5:

[0454] The server generates sentiment data based on the user's operation history and search history. It analyzes the user history data as input and calculates a sentiment score using a sentiment engine. It generates sentiment data as output and incorporates it into the next process. Specifically, this involves analyzing the user's behavior log and applying the sentiment scoring model.

[0455] Step 6:

[0456] The server adjusts the document's tags based on the generated sentiment data. It optimizes the tags using the tag list and sentiment data as input. It generates an adjusted tag list as output and re-assigns the tags to the document. Specifically, this involves adjusting the weighting of tags and modifying the tagging based on sentiment.

[0457] Step 7:

[0458] The terminal displays the generated tag list so that the user can review it. It displays the adjusted tag list as input and allows the user to provide feedback on the tags. It receives user feedback data as output. Specifically, it generates the tag list display interface and provides user input fields.

[0459] Step 8:

[0460] The server categorizes documents based on confirmed tags and sentiment data. It uses the final tag list and sentiment data as input to categorize documents. The categorization information is saved to the document database as output. Specifically, the process involves mapping tags to sentiment data and applying a categorization algorithm.

[0461] Step 9:

[0462] When a user performs a search using a specific keyword, the device sends the entered keyword to the server. Upon receiving the search keyword as input, the server searches its database for relevant documents. It then generates a list of search results and sends it back to the device. Specific operations include keyword matching, ranking of search results, and optimization of results based on sentiment data.

[0463] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0464] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0465] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0466] [Second Embodiment]

[0467] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0468] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0469] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0470] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0471] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0472] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0473] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0474] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0475] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0476] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0477] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0478] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0479] This invention relates to a system for automatically sorting and efficiently searching documents. This invention enables consistent document management that is not influenced by human subjectivity, and significantly improves search efficiency.

[0480] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[0481] Specific processing of the system

[0482] 1. User uploads documents

[0483] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[0484] 2. Server-driven summary generation

[0485] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, it uses summarization algorithms such as TextRank and BERT to shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0486] 3. Keyword extraction by the server

[0487] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[0488] 4. Tag generation by the server

[0489] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[0490] 5. Displaying tags on the device

[0491] The device displays a list of generated tags so that the user can review them. The user can then review the displayed tags and make any necessary corrections.

[0492] 6. Server-based document classification

[0493] The server categorizes documents based on the identified tags. Specifically, documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information." This information is stored in a database and used for later searches.

[0494] 7. User search and browsing

[0495] The user performs a search using specific keywords. The terminal sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the terminal, allowing the user to quickly find and view the documents they need.

[0496] Specific example

[0497] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[0498] summary:

[0499] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0500] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[0501] Users can review and modify these tags. Ultimately, the server categorizes documents based on these tags and stores them in the database. When a user searches using keywords such as "new customer," the server can quickly find relevant documents and display the results.

[0502] This allows for efficient document searching and management, without relying on human subjectivity.

[0503] The following describes the processing flow.

[0504] Step 1:

[0505] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[0506] Step 2:

[0507] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[0508] Step 3:

[0509] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[0510] Step 4:

[0511] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[0512] Step 5:

[0513] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[0514] Step 6:

[0515] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[0516] Step 7:

[0517] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[0518] Step 8:

[0519] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[0520] Step 9:

[0521] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[0522] Step 10:

[0523] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[0524] Step 11:

[0525] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[0526] Step 12:

[0527] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[0528] Step 13:

[0529] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[0530] Step 14:

[0531] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[0532] Step 15:

[0533] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[0534] Step 16:

[0535] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[0536] (Example 1)

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

[0538] Currently, many companies and individuals face the challenge of managing vast amounts of documents. In this situation, document classification and searching are often done manually, which is time-consuming and labor-intensive. Furthermore, management based on human subjectivity can lead to problems with the consistency of classification and the accuracy of searches. Therefore, there is a growing need for systems that can automatically organize documents and efficiently search them.

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

[0540] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for displaying the tags on a terminal, means for classifying documents based on the tags, and means for searching for documents classified based on the searched keywords. This enables automatic organization of documents, consistent document management, and efficient searching.

[0541] "Means of receiving documents" refers to the interface and communication means for users to upload document files to the server.

[0542] A "means for generating summaries from documents" refers to a summarization algorithm that uses natural language processing technology to summarize the content of a document, extract only the important information, and condense it into short sentences.

[0543] "Methods for extracting important keywords from summaries" refer to algorithms and processes for selecting frequently occurring words and semantically important words from generated summaries.

[0544] "Means for tagging documents based on extracted keywords" refers to algorithms and systems that automatically generate relevant tags based on extracted keywords.

[0545] "Means of displaying tags on a device" refers to a display screen and interface that allows users to check and modify the generated tags.

[0546] "Means for classifying documents based on tags" refers to algorithms and database management systems that automatically classify documents into appropriate categories using established tags.

[0547] "Means for searching for documents classified based on searched keywords" refers to a system and process that searches a database for relevant documents based on keywords entered by the user and displays the results.

[0548] This invention relates to a system for automatically sorting and efficiently searching documents. To implement this system, the following specific hardware and software are used to perform each processing step.

[0549] First, users utilize a dedicated web interface or application (e.g., a web browser) on a device such as a personal computer or smartphone. Through this, users can upload document files they wish to sort to the server. Specifically, using a common web browser such as Google Chrome or Mozilla Firefox, users select the documents they want to upload from a file selection dialog and click the "Upload" button.

[0550] The server receives document files sent by users and generates summaries of those documents using natural language processing (NLP) techniques. This process applies summarization algorithms such as TextRank and BERT. The server uses these algorithms to extract and analyze the key content of the document and create a concise summary.

[0551] Next, the server extracts key keywords from the generated summary. This process utilizes algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. This selects frequently occurring or semantically important words from the summary as keywords.

[0552] The server generates tags related to the document based on the extracted keywords. For example, if a sales report is uploaded, the server generates a summary such as "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this, it extracts keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies." Then, it generates tags such as "fiscal year 2023," "sales increase," and "new customers" from these keywords and attaches them to the document.

[0553] The device displays the generated tag list to the user. Through this display, the user can review the tags and modify or add them as needed. The display interface is intuitive and user-friendly, designed for easy operation.

[0554] The server then categorizes the documents based on the identified tags. This process utilizes a tag classification algorithm and a database management system. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[0555] Ultimately, when a user searches for a document using specific keywords, the terminal sends the entered keywords to the server. The server searches the database for the relevant documents and displays the search results on the terminal. This allows the user to quickly find and view the documents they need.

[0556] An example of a specific prompt message is as follows:

[0557] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0558] Please extract the key keywords from the summary above.

[0559] Please generate tags appropriate for the document in question.

[0560] Based on this, please categorize the document as a "sales report".

[0561] As described above, this system enables automatic document sorting and efficient searching. This allows for consistent document management and significantly improves search efficiency.

[0562] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0563] Step 1:

[0564] User upload of documents

[0565] Users upload document files using a dedicated web interface or application via a device such as a personal computer or smartphone. The user launches a browser, selects the document file from the file selection dialog, and clicks the "Upload" button.

[0566] Input: Document file selected by the user

[0567] Output: Document file sent to the server

[0568] Step 2:

[0569] Receiving documents from the server

[0570] The server receives document files sent by users. The received files are temporarily stored for subsequent processing.

[0571] Input: Uploaded document file

[0572] Output: Saved document file

[0573] Step 3:

[0574] Server-based summary generation

[0575] The server analyzes stored document files and generates summaries using natural language processing techniques. Specific algorithms used include TextRank and BERT. The server analyzes the document content, extracts key information, and generates a concise summary.

[0576] Input: Saved document file

[0577] Data processing: Summarization using natural language processing techniques (TextRank and BERT)

[0578] Output: Generated summary

[0579] Step 4:

[0580] Keyword extraction by the server

[0581] The server extracts key keywords from the generated summary. This process uses tools such as TF-IDF, Word2Vec, and SpaCy. It calculates the frequency and importance of words included in the summary and selects keywords.

[0582] Input: Generated summary

[0583] Data processing: Keyword extraction using TF-IDF, Word2Vec, and SpaCy.

[0584] Output: Extracted keyword list

[0585] Step 5:

[0586] Server-based tag generation

[0587] The server generates relevant tags based on the extracted keywords. It analyzes the keywords and creates tags based on specific conditions.

[0588] Input: Extracted keyword list

[0589] Data processing: Keyword-based tag generation algorithm

[0590] Output: Generated tag list

[0591] Step 6:

[0592] Tag display by device

[0593] The terminal displays a list of tags sent from the server to the user. The user reviews the displayed tags and makes corrections or additions as needed.

[0594] Input: Tag list sent from the server

[0595] Output: Tag list displayed to the user

[0596] Step 7:

[0597] Server-based document classification

[0598] The server classifies documents into specific categories based on confirmed tags. It uses a tag classification algorithm to assign documents to the appropriate categories and saves them to the database.

[0599] Input: Confirmed tag list

[0600] Data processing: Tag-based document classification algorithm

[0601] Output: Category information of documents stored in the database

[0602] Step 8:

[0603] User search and browsing

[0604] The user enters a specific keyword into the search bar on their device to search for documents. The device sends the entered keyword to the server, which searches its database for the relevant documents. The search results are displayed on the device, allowing the user to view the necessary documents.

[0605] Input: Search keywords entered by the user

[0606] Output: List of documents displayed as search results

[0607] Specific operation: User enters keywords into the search bar → Terminal sends keywords to the server → Server searches for documents → Search results are displayed on the terminal.

[0608] (Application Example 1)

[0609] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0610] Traditional document management systems suffer from the drawback of requiring many manual and time-consuming operations for document uploading, summary generation, keyword extraction, tagging, classification, and searching. Furthermore, for use in field settings such as logistics centers, efficient and intuitive operation is required, necessitating a flexible system utilizing smart devices. Smart search methods, such as voice commands, are also in demand.

[0611] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0612] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents, means for displaying tags, means for searching for classified documents based on searched keywords, means for reading documents via a smart device and uploading them to the cloud, and means for displaying search results on the smart device's display. This enables the automation of the document management process and efficient and intuitive document management using smart devices.

[0613] "Means of receiving documents" refers to the functions or devices that allow users to upload documents.

[0614] "Means for generating summaries from documents" refers to functions or devices that shorten and concisely express the main content of long documents.

[0615] "Means for extracting important keywords from a summary" refers to a function or device that selects important words from a generated summary.

[0616] "Means of tagging documents based on extracted keywords" refers to functions or devices that automatically assign labels related to documents based on keywords.

[0617] "Means of classifying documents" refers to functions or devices that categorize documents into specific categories based on their content.

[0618] "Means of displaying tags" refers to functions or devices that visually present tags assigned to a document to the user.

[0619] "Means for searching for documents classified based on searched keywords" refers to functions or devices that find appropriate documents based on keywords entered by the user.

[0620] "Means of reading documents via smart devices and uploading them to the cloud" refers to functions or devices that use smart glasses or other devices to scan documents and send them to a cloud server via the internet.

[0621] "Means of displaying search results on a smart device's display" refers to functions or devices that visually present search results to the user on the screen of a device such as smart glasses.

[0622] To implement this invention, the following system configuration and specific processing are used. The embodiment combines a server, a smart device (e.g., smart glasses), and cloud storage.

[0623] The server forms the core of a system that automatically sorts and efficiently searches documents, and has the following functions:

[0624] 1. Means of receiving documents:

[0625] Users use the smart glasses' camera to photograph documents, and the captured document data is uploaded to the cloud via the internet. The application on the smart glasses provides an interface for easily capturing documents and sending them to the server.

[0626] 2. Means for generating summaries from documents:

[0627] The server processes the received document and generates a summary using natural language processing techniques (e.g., TextRank or BERT). This summarization process shortens and concisely displays the main points of the document. For example, if a logistics instruction sheet is uploaded, a summary such as "Inventory list for Q2 2023 increased by 10% year-on-year" might be generated.

[0628] 3. Methods for extracting key keywords from summaries:

[0629] Next, the server extracts key keywords from the generated summary. This involves using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy) to identify and extract important concepts within the document.

[0630] 4. A means of tagging documents based on extracted keywords:

[0631] The system automatically generates relevant tags from the extracted keywords and adds them to the document. For example, tags such as "FY2023," "Q2," and "Inventory Increase" will be generated.

[0632] 5. Means of classifying documents:

[0633] The server categorizes documents based on the identified tags. This assigns documents to categories such as "periodical reports" and "inventory management." This information is stored in a database and used for later searches.

[0634] 6. Means of displaying tags:

[0635] The tag list generated by the server is displayed on the smart glasses' screen. The user can view the displayed tags and make corrections as needed.

[0636] 7. Means for searching for documents categorized based on searched keywords:

[0637] Users can use the voice command function of their smart glasses to perform searches using specific keywords. The smart glasses' display shows tags and summaries of relevant documents, allowing users to quickly find the documents they are looking for.

[0638] Hardware and software to use

[0639] Hardware: Smart glasses (e.g., Google Glass)

[0640] software:

[0641] Natural language processing models: spaCy, TextRank, BERT

[0642] Cloud storage: AWS S3, Google Cloud Storage

[0643] Specific example

[0644] When logistics center staff receive a new delivery order, they use smart glasses to photograph the document and upload it. The cloud server summarizes the document and generates tags. When they say the voice command "Search for new delivery orders," the smart glasses screen displays a list of relevant documents.

[0645] Example of a prompt

[0646] "Shipping Instructions: Generate a summary inventory list for Q2 2023 and extract tags. Output: Inventory List Summary: Inventory volume increased by 10% year-on-year in Q2 2023. Related tags: 2023, Q2, Inventory volume, 10% increase"

[0647] In this way, the document management process can be automated and streamlined, supporting the operations of the logistics center. This improves work efficiency and allows users to quickly find the documents they need.

[0648] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0649] Step 1:

[0650] The user uses smart glasses to photograph a document. This document data becomes the input. The user uses the smart glasses application to easily capture the document, and the captured document is uploaded to the cloud via the internet. The output of this process is the document data stored in cloud storage.

[0651] Step 2:

[0652] The server retrieves document data stored in cloud storage. This document data becomes the input, and the server uses natural language processing techniques (e.g., TextRank, BERT) to generate a summary of the document. The generated summary becomes the output. Specifically, it extracts the main content from the document and summarizes it in a shortened form.

[0653] Step 3:

[0654] The server extracts key keywords from the generated summary. This summary data becomes the input, and important concepts within the document are identified using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy). The extracted keywords become the output. Specifically, the frequency and context of the keywords are analyzed, and their importance is evaluated.

[0655] Step 4:

[0656] The server assigns tags to documents based on extracted keywords. This keyword data serves as input, and appropriate tags are automatically generated. The generated tag list is the output. Specifically, it searches the database for tags highly relevant to the keywords and assigns them to the documents.

[0657] Step 5:

[0658] The server categorizes documents. This tag list serves as input, and the server assigns documents to categories such as "periodical reports" and "inventory management." The category information of the classified documents is output. Specifically, it maps documents to the appropriate category based on their tags.

[0659] Step 6:

[0660] The server sends the generated tag list to the terminal, which displays it on the smart glasses' screen. This tag list serves as input, and the tags are displayed so that the user can see them. The displayed tag list becomes the output. Specifically, the tags are displayed on the smart glasses' HUD (Heads-Up Display).

[0661] Step 7:

[0662] The user performs a search using a specific keyword via voice command. This keyword becomes the input, and the server searches the database for relevant documents. The searched document list is the output. Specifically, the system quickly finds documents that match or are related to the search keyword from the database and displays the results on the smart glasses' display.

[0663] Through the above processes, users can efficiently manage documents using smart devices. Furthermore, the search function allows them to quickly find the documents they need.

[0664] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0665] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it is an invention that combines a sentiment engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results.

[0666] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[0667] Specific processing of the system

[0668] 1. User uploads documents

[0669] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[0670] 2. Server-driven summary generation

[0671] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, summarization algorithms such as TextRank and BERT are used to generate summaries that shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0672] 3. Keyword extraction by the server

[0673] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[0674] 4. Tag generation by the server

[0675] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[0676] 5. Acquisition of sentiment data by the server

[0677] The server is equipped with an emotion engine that generates emotion data from user input and operation history. For example, the emotion engine analyzes the user's emotion data based on information such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency.

[0678] 6. Server-side adjustment of tags and sentiment data

[0679] The server uses sentiment data generated by the sentiment engine to adjust document tagging and classification. For example, if a user is determined to be stressed based on their recent search history, the server will prioritize displaying documents related to relaxation.

[0680] 7. Displaying tags on the device

[0681] The device displays a list of generated tags so that the user can review them. The user can review the displayed tags and modify them as needed.

[0682] 8. Server-based document classification

[0683] The server categorizes documents based on established tags and sentiment data. Specifically, documents are assigned to categories such as "Sales Reports," "Annual," and "Customer Information." This information is stored in a database and used for later searches.

[0684] 9. User search and browsing

[0685] The user performs a search using specific keywords. The device sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the device, allowing the user to quickly find and view the documents they need. During this process, an emotion engine customizes the search results based on the user's emotions, prioritizing documents that meet the user's needs.

[0686] Specific example

[0687] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[0688] summary:

[0689] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0690] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[0691] The server is equipped with an emotion engine that generates emotion data from the user's operation history and search history. Based on this emotion data, tagging and document classification are adjusted.

[0692] Ultimately, when a user performs a search using keywords such as "new customer," the server can quickly search for relevant documents, taking sentiment data into consideration, and display the results. In this way, it becomes possible to efficiently search and manage documents without relying on human subjectivity.

[0693] The following describes the processing flow.

[0694] Step 1:

[0695] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[0696] Step 2:

[0697] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[0698] Step 3:

[0699] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[0700] Step 4:

[0701] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[0702] Step 5:

[0703] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[0704] Step 6:

[0705] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[0706] Step 7:

[0707] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[0708] Step 8:

[0709] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[0710] Step 9:

[0711] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[0712] Step 10:

[0713] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[0714] Step 11:

[0715] The server uses an emotion engine to generate emotion data from the user's operation history and search history. The emotion engine determines the user's emotional state and stores that data.

[0716] Step 12:

[0717] The server adjusts tagging and document classification based on emotional data. For example, if a user is feeling stressed, it will prioritize displaying relaxation-related documents.

[0718] Step 13:

[0719] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[0720] Step 14:

[0721] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[0722] Step 15:

[0723] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[0724] Step 16:

[0725] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[0726] Step 17:

[0727] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[0728] Step 18:

[0729] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[0730] Step 19:

[0731] The emotion engine customizes search results based on the user's emotional data. For example, if the emotion engine determines that the user is stressed, relaxation-related documents will be displayed preferentially.

[0732] (Example 2)

[0733] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0734] Conventional document management systems statically tag and classify documents, failing to reflect user emotions and intentions. As a result, it is difficult for users to quickly and accurately search for the documents they are looking for. Furthermore, there is a lack of methods to customize document search results and improve the user experience by utilizing sentiment data. This invention aims to achieve more efficient and customized document management and retrieval by dynamically adjusting document tagging and classification using user sentiment data.

[0735] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0736] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for acquiring user sentiment data, means for adjusting the classification of tags and documents based on the acquired sentiment data, means for classifying documents, means for displaying tags, and means for searching for classified documents based on searched keywords. This enables dynamic tagging and document classification based on user sentiment data.

[0737] "Means of receiving documents" refers to a system that sends documents uploaded by users to a server and has the function of receiving those documents on the server.

[0738] "Means for generating summaries from documents" refers to a function in which a server analyzes the entire content of a document and automatically generates a concise summary that highlights the main points.

[0739] "Method for extracting important keywords from a summary" refers to a function where the server analyzes the generated summary and extracts important words and phrases that represent the content of the document.

[0740] "A means of tagging documents based on extracted keywords" refers to a system that analyzes keywords and automatically assigns highly relevant tags to documents based on those keywords.

[0741] "Means for acquiring user sentiment data" refers to a function that analyzes the user's operation history and search history to collect data for inferring the user's emotional state.

[0742] "Means for adjusting tags and document classifications based on acquired sentiment data" refers to a system that dynamically changes document tags and classifications using collected sentiment data, performing appropriate tagging and document classification according to the user's emotional state.

[0743] "Means of classifying documents" refers to the function of dividing documents into specific categories or folders based on tags assigned to them.

[0744] "Means for displaying tags" refers to a function that visually displays tags generated by the server on the user interface, allowing users to review and modify them.

[0745] "Means for searching for documents classified based on searched keywords" refers to a function in which the server searches the database for relevant documents based on the search keywords entered by the user and displays the results.

[0746] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it incorporates an emotion engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results. The specific implementation method of this system is described below.

[0747] Hardware and software used

[0748] The server is the entity responsible for storing, analyzing, tagging, sentiment analysis, and retrieving document data. Typical software includes libraries and frameworks for natural language processing (NLP) (e.g., TextRank, BERT, SpaCy), algorithms for keyword extraction such as TF-IDF and Word2Vec, and sentiment analysis engines.

[0749] A terminal is a device used by users to upload documents, check tags, and receive search results. A common web browser or a dedicated application is used as the interface.

[0750] Users are those who utilize the system and perform operations such as uploading documents, checking and modifying tags, and searching.

[0751] Process Overview

[0752] Upload document

[0753] Users upload documents and files they want to sort to the server via their device. They can easily select files using a dedicated web interface or application and send them to the server by clicking the upload button.

[0754] Summary generation

[0755] The server analyzes the received document and generates a summary using natural language processing techniques such as TextRank and BERT. This summary is used to concisely grasp the main points of the document.

[0756] Keyword extraction

[0757] The server extracts important keywords from the generated summary using algorithms such as TF-IDF, Word2Vec, and SpaCy. This extracts words and phrases that represent the content of the document.

[0758] Tag generation

[0759] The server generates tags related to the document based on the extracted keywords. Using a tag generation algorithm, it converts the extracted keywords into tags and assigns them to the document.

[0760] Acquisition of emotional data

[0761] The server is equipped with an emotion engine that acquires emotion data based on the user's operation history and search history. The emotion engine analyzes the user's emotions from data such as what kind of searches the user performed and which documents they viewed.

[0762] Adjusting tags and document classification

[0763] The server adjusts document tagging and classification based on acquired emotional data. If the server determines that the user is stressed, it prioritizes displaying documents with tags and classifications related to relaxation.

[0764] Tag display

[0765] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify them as needed.

[0766] Document classification

[0767] The server categorizes documents based on established tags and sentiment data. Documents are categorized into categories such as "Sales Reports," "Annual," and "Customer Information," and stored in the database.

[0768] Search and browsing

[0769] The user enters specific keywords to perform a search. The server searches the database for relevant documents and sends the results to the user's terminal. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[0770] Specific example

[0771] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this summary, keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies" are extracted, and tags such as "fiscal year 2023," "sales increase," and "new customers" are generated.

[0772] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is seeking relaxation and will prioritize displaying related documents.

[0773] Examples of prompts for generative AI models

[0774] "After uploading the Sales Report.pdf file, perform text summarization, keyword extraction, and tag generation, then adjust the tags and document classification using sentiment data."

[0775] Thus, the system of the present invention can significantly improve the efficiency of document management and retrieval by utilizing user sentiment data.

[0776] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0777] Step 1:

[0778] User upload of documents

[0779] Users upload documents and files they want to categorize to the server using their device. Users open a dedicated web interface or application, click the file selection button, and select files such as "Sales Report.pdf". The selected files are sent from the device to the server by clicking the upload button.

[0780] Input: A document file uploaded by the user.

[0781] Output: Document file sent to the server.

[0782] Step 2:

[0783] Server-based summary generation

[0784] The server receives the uploaded document and determines the file format (PDF, Word, text, etc.). It extracts the document content using an appropriate parser and generates a document summary using natural language processing techniques (TextRank or BERT). The server analyzes the document content and creates a concise summary that includes the main points.

[0785] Input: Document file sent to the server.

[0786] Output: A summary of the document.

[0787] Specific example: The server analyzes "Sales Report.pdf" and generates a summary stating, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0788] Step 3:

[0789] Keyword extraction by the server

[0790] The server extracts key keywords from the generated summary. It uses algorithms such as TF-IDF, Word2Vec, and SpaCy to analyze the summary and select frequently occurring and important words.

[0791] Input: A summary of the document.

[0792] Output: A list of extracted important keywords.

[0793] Specific example: Extract keywords such as "FY2023," "First Quarter," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" from the summary.

[0794] Step 4:

[0795] Server-based tag generation

[0796] The server generates tags related to the document based on the extracted keywords. It converts the keywords into the appropriate tag format and assigns them to the document. A tag generation algorithm is used to automatically create relevant tags from the extracted keywords.

[0797] Input: A list of important keywords extracted.

[0798] Output: A list of tags assigned to the document.

[0799] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are generated.

[0800] Step 5:

[0801] Acquisition of emotional data by a server

[0802] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. It analyzes data such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency, in order to infer the user's emotional state.

[0803] Input: Data such as user activity history, search history, referenced documents, time of day, and frequency.

[0804] Output: Generated user sentiment data.

[0805] Specific example: If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is experiencing stress.

[0806] Step 6:

[0807] Server-based adjustment of tag and document classification

[0808] The server adjusts document tagging and classification based on the acquired sentiment data. It re-evaluates tags based on the sentiment data and makes changes or additions as needed.

[0809] Input: Generated user sentiment data, list of tags assigned to the document.

[0810] Output: Tag lists and categorized documents adjusted based on sentiment data.

[0811] Specific example: If a user is determined to be experiencing stress, documents with tags or categories related to relaxation will be prioritized for display.

[0812] Step 7:

[0813] Tag display by device

[0814] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify or add them as needed.

[0815] Input: A list of tags adjusted based on sentiment data.

[0816] Output: A list of tags displayed on the user's screen.

[0817] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are displayed on the user's screen.

[0818] Step 8:

[0819] Server-based document classification

[0820] The server categorizes documents based on established tags and sentiment data. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[0821] Input: Tag list adjusted based on sentiment data, document file.

[0822] Output: Category information for classified documents.

[0823] Specific example: "Sales Report.pdf" is categorized as "Sales Report," "Fiscal Year 2023," etc.

[0824] Step 9:

[0825] User search and browsing

[0826] The user enters specific keywords to perform a search. The terminal sends the entered keywords to the server, which searches its database for relevant documents. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[0827] Input: Search keywords entered by the user.

[0828] Output: Customized search results list.

[0829] Specific example: When a user searches using the keyword "new customer," relevant documents are displayed preferentially, and "Sales Report.pdf" is found among them.

[0830] (Application Example 2)

[0831] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0832] In today's information society, a vast amount of documents are stored electronically, making it difficult to efficiently search for necessary documents due to their sheer volume. Furthermore, conventional document management systems rely solely on simple keyword matching without considering the user's emotional state, making it difficult to quickly provide the information the user truly needs. Therefore, there is a need for a system that utilizes user emotional data to customize search results and provide a more personalized search experience.

[0833] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0834] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents using the tags, means for generating user sentiment data, means for adjusting document tags based on sentiment data, means for displaying the generated tags, and means for searching for documents classified based on searched keywords and adjusted tags. This enables automatic document classification and customization of search results based on user sentiment.

[0835] "Means of receiving documents" refers to a function that allows users to upload document data to a server via their device.

[0836] "Means for generating summaries from documents" refers to functions that use natural language processing technology to shorten and concisely display the main content of a document.

[0837] "Methods for extracting important keywords" refer to functions that extract frequently occurring and important information from document summaries.

[0838] "Means of tagging documents" refers to a function that automatically generates and assigns tags related to a document based on extracted keywords.

[0839] "A means of classifying documents using tags" refers to a function that sorts documents into specific categories based on the tags that have been generated.

[0840] "Means for generating user emotion data" refers to a function that analyzes a user's search history and operation history and generates data on their emotional state.

[0841] "Means for adjusting document tags based on sentiment data" refers to a function that dynamically changes the tags and classifications assigned to a document based on generated sentiment data.

[0842] "Means for displaying generated tags" refers to a function that displays a list of tags so that users can review and modify them.

[0843] "Means for searching for documents categorized based on searched keywords and adjusted tags" refers to a function that searches a database for appropriate documents based on keywords and adjusted tags entered by the user and displays the results.

[0844] The system that realizes this invention is designed to allow users to search for documents in an efficient and personalized manner. Specific embodiments are described below.

[0845] Hardware and software to be used

[0846] This system is recommended to use the following hardware and software.

[0847] Hardware:

[0848] CPU: Intel i7 or higher recommended

[0849] Memory: 16GB or more

[0850] Storage: 500GB SSD

[0851] User devices: PCs, smartphones, tablets, etc.

[0852] software:

[0853] OS: Windows 10 or Linux Ubuntu 20.04

[0854] Libraries: SpaCy (en_core_web_sm model), TextBlob, Python 3.x

[0855] System configuration and operation

[0856] 1. Upload document

[0857] Users upload documents and files they want to categorize to the server via their device.

[0858] A dedicated web interface or smartphone app can be used.

[0859] 2. Summary generation

[0860] The server receives the uploaded document and generates a summary using natural language processing technology (such as SpaCy).

[0861] For example, summarization algorithms such as TextRank and BERT are used to shorten and concisely display the main content of a document.

[0862] 3. Keyword Extraction

[0863] The process of extracting key keywords from the summary uses TF-IDF, Word2Vec, and SpaCy.

[0864] For example, keywords such as "sales revenue" and "new customers" can be extracted from the content of uploaded documents.

[0865] 4. Tag generation

[0866] Based on the extracted keywords, generate tags relevant to the document.

[0867] This automatically adds tags such as "increased sales" and "new customers" to documents.

[0868] 5. Generation and adjustment of emotional data

[0869] The server generates sentiment data using a sentiment engine based on the user's operation history and search history.

[0870] For example, the document's tags are adjusted based on whether the user has recently viewed many negative reviews.

[0871] 6. Displaying tags and classifying documents

[0872] The device displays the generated tag list so that the user can review it and make corrections as needed.

[0873] The server categorizes documents into specific categories based on confirmed tags and sentiment data.

[0874] 7. Search and Browsing

[0875] When a user performs a search using a specific keyword, the device sends that keyword to the server.

[0876] The server searches the database for relevant documents and displays customized search results based on sentiment data.

[0877] Specific example

[0878] For example, if a user uploads a "Sales Report for the First Quarter of Fiscal Year 2023," the server receives the document, generates a summary, and displays a summary like the following:

[0879] summary:

[0880] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0881] Furthermore, keywords such as "FY2023," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" are extracted from the summary, and tags are generated based on these. After tag generation, if the user has recently seen many positive reviews, a "Positive" tag is added, adjusting the search results according to the user's sentiment.

[0882] Example of a prompt

[0883] Please write the following program to generate a user sentiment score based on reviews and adjust the document tags accordingly.

[0884] 1. Use Spasi to summarize the document.

[0885] 2. Extract keywords from the summary.

[0886] 3. Tags are generated based on the keywords extracted by Spasi.

[0887] 4. Calculate sentiment data from user reviews. (Examples: 'This product is amazing!', 'Not satisfied', 'Excellent quality')

[0888] 5. Adjust the tags generated based on the calculated sentiment data. (Add a "negative" tag if the sentiment is negative, and a "positive" tag if it's positive.)

[0889] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0890] Step 1:

[0891] The server receives documents uploaded by users via their terminals. The server receives the document data (PDF, DOCX, TXT, etc.) as input and stores it within the document management system. At the same time, the document's metadata (title, creation date, etc.) is also saved.

[0892] Step 2:

[0893] The server analyzes the received document data using natural language processing techniques (such as SpaCy) and generates a document summary. It analyzes the document text as input, extracts the main content, and generates a summary. It generates a summary text as output and passes it on to the next process. Specifically, this involves sentence segmentation, importance calculation, and selection of summary sentences.

[0894] Step 3:

[0895] The server extracts key keywords from the generated summary. Based on the summary text as input, it selects key keywords using algorithms such as TF-IDF and Word2Vec. It then generates a list of key keywords as output. Specifically, this involves calculating word frequency, applying inverse document frequency, and calculating weights.

[0896] Step 4:

[0897] The server automatically generates and assigns tags related to the document based on the extracted keywords. It uses a keyword list as input to select tags to be assigned to the document. It generates a tag list as output and associates it with the document. Specifically, this involves keyword mapping and the generation of category tags.

[0898] Step 5:

[0899] The server generates sentiment data based on the user's operation history and search history. It analyzes the user history data as input and calculates a sentiment score using a sentiment engine. It generates sentiment data as output and incorporates it into the next process. Specifically, this involves analyzing the user's behavior log and applying the sentiment scoring model.

[0900] Step 6:

[0901] The server adjusts the document's tags based on the generated sentiment data. It optimizes the tags using the tag list and sentiment data as input. It generates an adjusted tag list as output and re-assigns the tags to the document. Specifically, this involves adjusting the weighting of tags and modifying the tagging based on sentiment.

[0902] Step 7:

[0903] The terminal displays the generated tag list so that the user can review it. It displays the adjusted tag list as input and allows the user to provide feedback on the tags. It receives user feedback data as output. Specifically, it generates the tag list display interface and provides user input fields.

[0904] Step 8:

[0905] The server categorizes documents based on confirmed tags and sentiment data. It uses the final tag list and sentiment data as input to categorize documents. The categorization information is saved to the document database as output. Specifically, the process involves mapping tags to sentiment data and applying a categorization algorithm.

[0906] Step 9:

[0907] When a user performs a search using a specific keyword, the device sends the entered keyword to the server. Upon receiving the search keyword as input, the server searches its database for relevant documents. It then generates a list of search results and sends it back to the device. Specific operations include keyword matching, ranking of search results, and optimization of results based on sentiment data.

[0908] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0909] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0910] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0911] [Third Embodiment]

[0912] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0913] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0914] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0915] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0916] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0917] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0918] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0919] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0920] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0921] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0922] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0923] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0924] This invention relates to a system for automatically sorting and efficiently searching documents. This invention enables consistent document management that is not influenced by human subjectivity, and significantly improves search efficiency.

[0925] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[0926] Specific processing of the system

[0927] 1. User uploads documents

[0928] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[0929] 2. Server-driven summary generation

[0930] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, it uses summarization algorithms such as TextRank and BERT to shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[0931] 3. Keyword extraction by the server

[0932] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[0933] 4. Tag generation by the server

[0934] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[0935] 5. Displaying tags on the device

[0936] The device displays a list of generated tags so that the user can review them. The user can then review the displayed tags and make any necessary corrections.

[0937] 6. Server-based document classification

[0938] The server categorizes documents based on the identified tags. Specifically, documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information." This information is stored in a database and used for later searches.

[0939] 7. User search and browsing

[0940] The user performs a search using specific keywords. The terminal sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the terminal, allowing the user to quickly find and view the documents they need.

[0941] Specific example

[0942] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[0943] summary:

[0944] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[0945] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[0946] Users can review and modify these tags. Ultimately, the server categorizes documents based on these tags and stores them in the database. When a user searches using keywords such as "new customer," the server can quickly find relevant documents and display the results.

[0947] This allows for efficient document searching and management, without relying on human subjectivity.

[0948] The following describes the processing flow.

[0949] Step 1:

[0950] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[0951] Step 2:

[0952] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[0953] Step 3:

[0954] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[0955] Step 4:

[0956] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[0957] Step 5:

[0958] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[0959] Step 6:

[0960] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[0961] Step 7:

[0962] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[0963] Step 8:

[0964] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[0965] Step 9:

[0966] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[0967] Step 10:

[0968] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[0969] Step 11:

[0970] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[0971] Step 12:

[0972] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[0973] Step 13:

[0974] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[0975] Step 14:

[0976] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[0977] Step 15:

[0978] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[0979] Step 16:

[0980] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[0981] (Example 1)

[0982] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0983] Currently, many companies and individuals face the challenge of managing vast amounts of documents. In this situation, document classification and searching are often done manually, which is time-consuming and labor-intensive. Furthermore, management based on human subjectivity can lead to problems with the consistency of classification and the accuracy of searches. Therefore, there is a growing need for systems that can automatically organize documents and efficiently search them.

[0984] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0985] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for displaying the tags on a terminal, means for classifying documents based on the tags, and means for searching for documents classified based on the searched keywords. This enables automatic organization of documents, consistent document management, and efficient searching.

[0986] "Means of receiving documents" refers to the interface and communication means for users to upload document files to the server.

[0987] A "means for generating summaries from documents" refers to a summarization algorithm that uses natural language processing technology to summarize the content of a document, extract only the important information, and condense it into short sentences.

[0988] "Methods for extracting important keywords from summaries" refer to algorithms and processes for selecting frequently occurring words and semantically important words from generated summaries.

[0989] "Means for tagging documents based on extracted keywords" refers to algorithms and systems that automatically generate relevant tags based on extracted keywords.

[0990] "Means of displaying tags on a device" refers to a display screen and interface that allows users to check and modify the generated tags.

[0991] "Means for classifying documents based on tags" refers to algorithms and database management systems that automatically classify documents into appropriate categories using established tags.

[0992] "Means for searching for documents classified based on searched keywords" refers to a system and process that searches a database for relevant documents based on keywords entered by the user and displays the results.

[0993] This invention relates to a system for automatically sorting and efficiently searching documents. To implement this system, the following specific hardware and software are used to perform each processing step.

[0994] First, users utilize a dedicated web interface or application (e.g., a web browser) on a device such as a personal computer or smartphone. Through this, users can upload document files they wish to sort to the server. Specifically, using a common web browser such as Google Chrome or Mozilla Firefox, users select the documents they want to upload from a file selection dialog and click the "Upload" button.

[0995] The server receives document files sent by users and generates summaries of those documents using natural language processing (NLP) techniques. This process applies summarization algorithms such as TextRank and BERT. The server uses these algorithms to extract and analyze the key content of the document and create a concise summary.

[0996] Next, the server extracts key keywords from the generated summary. This process utilizes algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. This selects frequently occurring or semantically important words from the summary as keywords.

[0997] The server generates tags related to the document based on the extracted keywords. For example, if a sales report is uploaded, the server generates a summary such as "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this, it extracts keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies." Then, it generates tags such as "fiscal year 2023," "sales increase," and "new customers" from these keywords and attaches them to the document.

[0998] The device displays the generated tag list to the user. Through this display, the user can review the tags and modify or add them as needed. The display interface is intuitive and user-friendly, designed for easy operation.

[0999] The server then categorizes the documents based on the identified tags. This process utilizes a tag classification algorithm and a database management system. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[1000] Ultimately, when a user searches for a document using specific keywords, the terminal sends the entered keywords to the server. The server searches the database for the relevant documents and displays the search results on the terminal. This allows the user to quickly find and view the documents they need.

[1001] An example of a specific prompt message is as follows:

[1002] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1003] Please extract the key keywords from the summary above.

[1004] Please generate tags appropriate for the document in question.

[1005] Based on this, please categorize the document as a "sales report".

[1006] As described above, this system enables automatic document sorting and efficient searching. This allows for consistent document management and significantly improves search efficiency.

[1007] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1008] Step 1:

[1009] User upload of documents

[1010] Users upload document files using a dedicated web interface or application via a device such as a personal computer or smartphone. The user launches a browser, selects the document file from the file selection dialog, and clicks the "Upload" button.

[1011] Input: Document file selected by the user

[1012] Output: Document file sent to the server

[1013] Step 2:

[1014] Receiving documents from the server

[1015] The server receives document files sent by users. The received files are temporarily stored for subsequent processing.

[1016] Input: Uploaded document file

[1017] Output: Saved document file

[1018] Step 3:

[1019] Server-based summary generation

[1020] The server analyzes stored document files and generates summaries using natural language processing techniques. Specific algorithms used include TextRank and BERT. The server analyzes the document content, extracts key information, and generates a concise summary.

[1021] Input: Saved document file

[1022] Data processing: Summarization using natural language processing techniques (TextRank and BERT)

[1023] Output: Generated summary

[1024] Step 4:

[1025] Keyword extraction by the server

[1026] The server extracts key keywords from the generated summary. This process uses tools such as TF-IDF, Word2Vec, and SpaCy. It calculates the frequency and importance of words included in the summary and selects keywords.

[1027] Input: Generated summary

[1028] Data processing: Keyword extraction using TF-IDF, Word2Vec, and SpaCy.

[1029] Output: Extracted keyword list

[1030] Step 5:

[1031] Server-based tag generation

[1032] The server generates relevant tags based on the extracted keywords. It analyzes the keywords and creates tags based on specific conditions.

[1033] Input: Extracted keyword list

[1034] Data processing: Keyword-based tag generation algorithm

[1035] Output: Generated tag list

[1036] Step 6:

[1037] Tag display by device

[1038] The terminal displays a list of tags sent from the server to the user. The user reviews the displayed tags and makes corrections or additions as needed.

[1039] Input: Tag list sent from the server

[1040] Output: Tag list displayed to the user

[1041] Step 7:

[1042] Server-based document classification

[1043] The server classifies documents into specific categories based on confirmed tags. It uses a tag classification algorithm to assign documents to the appropriate categories and saves them to the database.

[1044] Input: Confirmed tag list

[1045] Data processing: Tag-based document classification algorithm

[1046] Output: Category information of documents stored in the database

[1047] Step 8:

[1048] User search and browsing

[1049] The user enters a specific keyword into the search bar on their device to search for documents. The device sends the entered keyword to the server, which searches its database for the relevant documents. The search results are displayed on the device, allowing the user to view the necessary documents.

[1050] Input: Search keywords entered by the user

[1051] Output: List of documents displayed as search results

[1052] Specific operation: User enters keywords into the search bar → Terminal sends keywords to the server → Server searches for documents → Search results are displayed on the terminal.

[1053] (Application Example 1)

[1054] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1055] Traditional document management systems suffer from the drawback of requiring many manual and time-consuming operations for document uploading, summary generation, keyword extraction, tagging, classification, and searching. Furthermore, for use in field settings such as logistics centers, efficient and intuitive operation is required, necessitating a flexible system utilizing smart devices. Smart search methods, such as voice commands, are also in demand.

[1056] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1057] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents, means for displaying tags, means for searching for classified documents based on searched keywords, means for reading documents via a smart device and uploading them to the cloud, and means for displaying search results on the smart device's display. This enables the automation of the document management process and efficient and intuitive document management using smart devices.

[1058] "Means of receiving documents" refers to the functions or devices that allow users to upload documents.

[1059] "Means for generating summaries from documents" refers to functions or devices that shorten and concisely express the main content of long documents.

[1060] "Means for extracting important keywords from a summary" refers to a function or device that selects important words from a generated summary.

[1061] "Means of tagging documents based on extracted keywords" refers to functions or devices that automatically assign labels related to documents based on keywords.

[1062] "Means of classifying documents" refers to functions or devices that categorize documents into specific categories based on their content.

[1063] "Means of displaying tags" refers to functions or devices that visually present tags assigned to a document to the user.

[1064] "Means for searching for documents classified based on searched keywords" refers to functions or devices that find appropriate documents based on keywords entered by the user.

[1065] "Means of reading documents via smart devices and uploading them to the cloud" refers to functions or devices that use smart glasses or other devices to scan documents and send them to a cloud server via the internet.

[1066] "Means of displaying search results on a smart device's display" refers to functions or devices that visually present search results to the user on the screen of a device such as smart glasses.

[1067] To implement this invention, the following system configuration and specific processing are used. The embodiment combines a server, a smart device (e.g., smart glasses), and cloud storage.

[1068] The server forms the core of a system that automatically sorts and efficiently searches documents, and has the following functions:

[1069] 1. Means of receiving documents:

[1070] Users use the smart glasses' camera to photograph documents, and the captured document data is uploaded to the cloud via the internet. The application on the smart glasses provides an interface for easily capturing documents and sending them to the server.

[1071] 2. Means for generating summaries from documents:

[1072] The server processes the received document and generates a summary using natural language processing techniques (e.g., TextRank or BERT). This summarization process shortens and concisely displays the main points of the document. For example, if a logistics instruction sheet is uploaded, a summary such as "Inventory list for Q2 2023 increased by 10% year-on-year" might be generated.

[1073] 3. Methods for extracting key keywords from summaries:

[1074] Next, the server extracts key keywords from the generated summary. This involves using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy) to identify and extract important concepts within the document.

[1075] 4. A means of tagging documents based on extracted keywords:

[1076] The system automatically generates relevant tags from the extracted keywords and adds them to the document. For example, tags such as "FY2023," "Q2," and "Inventory Increase" will be generated.

[1077] 5. Means of classifying documents:

[1078] The server categorizes documents based on the identified tags. This assigns documents to categories such as "periodical reports" and "inventory management." This information is stored in a database and used for later searches.

[1079] 6. Means of displaying tags:

[1080] The tag list generated by the server is displayed on the smart glasses' screen. The user can view the displayed tags and make corrections as needed.

[1081] 7. Means for searching for documents categorized based on searched keywords:

[1082] Users can use the voice command function of their smart glasses to perform searches using specific keywords. The smart glasses' display shows tags and summaries of relevant documents, allowing users to quickly find the documents they are looking for.

[1083] Hardware and software to use

[1084] Hardware: Smart glasses (e.g., Google Glass)

[1085] software:

[1086] Natural language processing models: spaCy, TextRank, BERT

[1087] Cloud storage: AWS S3, Google Cloud Storage

[1088] Specific example

[1089] When logistics center staff receive a new delivery order, they use smart glasses to photograph the document and upload it. The cloud server summarizes the document and generates tags. When they say the voice command "Search for new delivery orders," the smart glasses screen displays a list of relevant documents.

[1090] Example of a prompt

[1091] "Shipping Instructions: Generate a summary inventory list for Q2 2023 and extract tags. Output: Inventory List Summary: Inventory volume increased by 10% year-on-year in Q2 2023. Related tags: 2023, Q2, Inventory volume, 10% increase"

[1092] In this way, the document management process can be automated and streamlined, supporting the operations of the logistics center. This improves work efficiency and allows users to quickly find the documents they need.

[1093] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1094] Step 1:

[1095] The user uses smart glasses to photograph a document. This document data becomes the input. The user uses the smart glasses application to easily capture the document, and the captured document is uploaded to the cloud via the internet. The output of this process is the document data stored in cloud storage.

[1096] Step 2:

[1097] The server retrieves document data stored in cloud storage. This document data becomes the input, and the server uses natural language processing techniques (e.g., TextRank, BERT) to generate a summary of the document. The generated summary becomes the output. Specifically, it extracts the main content from the document and summarizes it in a shortened form.

[1098] Step 3:

[1099] The server extracts key keywords from the generated summary. This summary data becomes the input, and important concepts within the document are identified using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy). The extracted keywords become the output. Specifically, the frequency and context of the keywords are analyzed, and their importance is evaluated.

[1100] Step 4:

[1101] The server assigns tags to documents based on extracted keywords. This keyword data serves as input, and appropriate tags are automatically generated. The generated tag list is the output. Specifically, it searches the database for tags highly relevant to the keywords and assigns them to the documents.

[1102] Step 5:

[1103] The server categorizes documents. This tag list serves as input, and the server assigns documents to categories such as "periodical reports" and "inventory management." The category information of the classified documents is output. Specifically, it maps documents to the appropriate category based on their tags.

[1104] Step 6:

[1105] The server sends the generated tag list to the terminal, which displays it on the smart glasses' screen. This tag list serves as input, and the tags are displayed so that the user can see them. The displayed tag list becomes the output. Specifically, the tags are displayed on the smart glasses' HUD (Heads-Up Display).

[1106] Step 7:

[1107] The user performs a search using a specific keyword via voice command. This keyword becomes the input, and the server searches the database for relevant documents. The searched document list is the output. Specifically, the system quickly finds documents that match or are related to the search keyword from the database and displays the results on the smart glasses' display.

[1108] Through the above processes, users can efficiently manage documents using smart devices. Furthermore, the search function allows them to quickly find the documents they need.

[1109] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1110] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it is an invention that combines a sentiment engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results.

[1111] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[1112] Specific processing of the system

[1113] 1. User uploads documents

[1114] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[1115] 2. Server-driven summary generation

[1116] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, summarization algorithms such as TextRank and BERT are used to generate summaries that shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[1117] 3. Keyword extraction by the server

[1118] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[1119] 4. Tag generation by the server

[1120] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[1121] 5. Acquisition of sentiment data by the server

[1122] The server is equipped with an emotion engine that generates emotion data from user input and operation history. For example, the emotion engine analyzes the user's emotion data based on information such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency.

[1123] 6. Server-side adjustment of tags and sentiment data

[1124] The server uses sentiment data generated by the sentiment engine to adjust document tagging and classification. For example, if a user is determined to be stressed based on their recent search history, the server will prioritize displaying documents related to relaxation.

[1125] 7. Displaying tags on the device

[1126] The device displays a list of generated tags so that the user can review them. The user can review the displayed tags and modify them as needed.

[1127] 8. Server-based document classification

[1128] The server categorizes documents based on established tags and sentiment data. Specifically, documents are assigned to categories such as "Sales Reports," "Annual," and "Customer Information." This information is stored in a database and used for later searches.

[1129] 9. User search and browsing

[1130] The user performs a search using specific keywords. The device sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the device, allowing the user to quickly find and view the documents they need. During this process, an emotion engine customizes the search results based on the user's emotions, prioritizing documents that meet the user's needs.

[1131] Specific example

[1132] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[1133] summary:

[1134] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1135] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[1136] The server is equipped with an emotion engine that generates emotion data from the user's operation history and search history. Based on this emotion data, tagging and document classification are adjusted.

[1137] Ultimately, when a user performs a search using keywords such as "new customer," the server can quickly search for relevant documents, taking sentiment data into consideration, and display the results. In this way, it becomes possible to efficiently search and manage documents without relying on human subjectivity.

[1138] The following describes the processing flow.

[1139] Step 1:

[1140] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[1141] Step 2:

[1142] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[1143] Step 3:

[1144] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[1145] Step 4:

[1146] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[1147] Step 5:

[1148] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[1149] Step 6:

[1150] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[1151] Step 7:

[1152] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[1153] Step 8:

[1154] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[1155] Step 9:

[1156] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[1157] Step 10:

[1158] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[1159] Step 11:

[1160] The server uses an emotion engine to generate emotion data from the user's operation history and search history. The emotion engine determines the user's emotional state and stores that data.

[1161] Step 12:

[1162] The server adjusts tagging and document classification based on emotional data. For example, if a user is feeling stressed, it will prioritize displaying relaxation-related documents.

[1163] Step 13:

[1164] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[1165] Step 14:

[1166] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[1167] Step 15:

[1168] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[1169] Step 16:

[1170] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[1171] Step 17:

[1172] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[1173] Step 18:

[1174] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[1175] Step 19:

[1176] The emotion engine customizes search results based on the user's emotional data. For example, if the emotion engine determines that the user is stressed, relaxation-related documents will be displayed preferentially.

[1177] (Example 2)

[1178] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1179] Conventional document management systems statically tag and classify documents, failing to reflect user emotions and intentions. As a result, it is difficult for users to quickly and accurately search for the documents they are looking for. Furthermore, there is a lack of methods to customize document search results and improve the user experience by utilizing sentiment data. This invention aims to achieve more efficient and customized document management and retrieval by dynamically adjusting document tagging and classification using user sentiment data.

[1180] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1181] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for acquiring user sentiment data, means for adjusting the classification of tags and documents based on the acquired sentiment data, means for classifying documents, means for displaying tags, and means for searching for classified documents based on searched keywords. This enables dynamic tagging and document classification based on user sentiment data.

[1182] "Means of receiving documents" refers to a system that sends documents uploaded by users to a server and has the function of receiving those documents on the server.

[1183] "Means for generating summaries from documents" refers to a function in which a server analyzes the entire content of a document and automatically generates a concise summary that highlights the main points.

[1184] "Method for extracting important keywords from a summary" refers to a function where the server analyzes the generated summary and extracts important words and phrases that represent the content of the document.

[1185] "A means of tagging documents based on extracted keywords" refers to a system that analyzes keywords and automatically assigns highly relevant tags to documents based on those keywords.

[1186] "Means for acquiring user sentiment data" refers to a function that analyzes the user's operation history and search history to collect data for inferring the user's emotional state.

[1187] "Means for adjusting tags and document classifications based on acquired sentiment data" refers to a system that dynamically changes document tags and classifications using collected sentiment data, performing appropriate tagging and document classification according to the user's emotional state.

[1188] "Means of classifying documents" refers to the function of dividing documents into specific categories or folders based on tags assigned to them.

[1189] "Means for displaying tags" refers to a function that visually displays tags generated by the server on the user interface, allowing users to review and modify them.

[1190] "Means for searching for documents classified based on searched keywords" refers to a function in which the server searches the database for relevant documents based on the search keywords entered by the user and displays the results.

[1191] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it incorporates an emotion engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results. The specific implementation method of this system is described below.

[1192] Hardware and software used

[1193] The server is the entity responsible for storing, analyzing, tagging, sentiment analysis, and retrieving document data. Typical software includes libraries and frameworks for natural language processing (NLP) (e.g., TextRank, BERT, SpaCy), algorithms for keyword extraction such as TF-IDF and Word2Vec, and sentiment analysis engines.

[1194] A terminal is a device used by users to upload documents, check tags, and receive search results. A common web browser or a dedicated application is used as the interface.

[1195] Users are those who utilize the system and perform operations such as uploading documents, checking and modifying tags, and searching.

[1196] Process Overview

[1197] Upload document

[1198] Users upload documents and files they want to sort to the server via their device. They can easily select files using a dedicated web interface or application and send them to the server by clicking the upload button.

[1199] Summary generation

[1200] The server analyzes the received document and generates a summary using natural language processing techniques such as TextRank and BERT. This summary is used to concisely grasp the main points of the document.

[1201] Keyword extraction

[1202] The server extracts important keywords from the generated summary using algorithms such as TF-IDF, Word2Vec, and SpaCy. This extracts words and phrases that represent the content of the document.

[1203] Tag generation

[1204] The server generates tags related to the document based on the extracted keywords. Using a tag generation algorithm, it converts the extracted keywords into tags and assigns them to the document.

[1205] Acquisition of emotional data

[1206] The server is equipped with an emotion engine that acquires emotion data based on the user's operation history and search history. The emotion engine analyzes the user's emotions from data such as what kind of searches the user performed and which documents they viewed.

[1207] Adjusting tags and document classification

[1208] The server adjusts document tagging and classification based on acquired emotional data. If the server determines that the user is stressed, it prioritizes displaying documents with tags and classifications related to relaxation.

[1209] Tag display

[1210] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify them as needed.

[1211] Document classification

[1212] The server categorizes documents based on established tags and sentiment data. Documents are categorized into categories such as "Sales Reports," "Annual," and "Customer Information," and stored in the database.

[1213] Search and browsing

[1214] The user enters specific keywords to perform a search. The server searches the database for relevant documents and sends the results to the user's terminal. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[1215] Specific example

[1216] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this summary, keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies" are extracted, and tags such as "fiscal year 2023," "sales increase," and "new customers" are generated.

[1217] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is seeking relaxation and will prioritize displaying related documents.

[1218] Examples of prompts for generative AI models

[1219] "After uploading the Sales Report.pdf file, perform text summarization, keyword extraction, and tag generation, then adjust the tags and document classification using sentiment data."

[1220] Thus, the system of the present invention can significantly improve the efficiency of document management and retrieval by utilizing user sentiment data.

[1221] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1222] Step 1:

[1223] User upload of documents

[1224] Users upload documents and files they want to categorize to the server using their device. Users open a dedicated web interface or application, click the file selection button, and select files such as "Sales Report.pdf". The selected files are sent from the device to the server by clicking the upload button.

[1225] Input: A document file uploaded by the user.

[1226] Output: Document file sent to the server.

[1227] Step 2:

[1228] Server-based summary generation

[1229] The server receives the uploaded document and determines the file format (PDF, Word, text, etc.). It extracts the document content using an appropriate parser and generates a document summary using natural language processing techniques (TextRank or BERT). The server analyzes the document content and creates a concise summary that includes the main points.

[1230] Input: Document file sent to the server.

[1231] Output: A summary of the document.

[1232] Specific example: The server analyzes "Sales Report.pdf" and generates a summary stating, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[1233] Step 3:

[1234] Keyword extraction by the server

[1235] The server extracts key keywords from the generated summary. It uses algorithms such as TF-IDF, Word2Vec, and SpaCy to analyze the summary and select frequently occurring and important words.

[1236] Input: A summary of the document.

[1237] Output: A list of extracted important keywords.

[1238] Specific example: Extract keywords such as "FY2023," "First Quarter," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" from the summary.

[1239] Step 4:

[1240] Server-based tag generation

[1241] The server generates tags related to the document based on the extracted keywords. It converts the keywords into the appropriate tag format and assigns them to the document. A tag generation algorithm is used to automatically create relevant tags from the extracted keywords.

[1242] Input: A list of important keywords extracted.

[1243] Output: A list of tags assigned to the document.

[1244] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are generated.

[1245] Step 5:

[1246] Acquisition of emotional data by a server

[1247] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. It analyzes data such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency, in order to infer the user's emotional state.

[1248] Input: Data such as user activity history, search history, referenced documents, time of day, and frequency.

[1249] Output: Generated user sentiment data.

[1250] Specific example: If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is experiencing stress.

[1251] Step 6:

[1252] Server-based adjustment of tag and document classification

[1253] The server adjusts document tagging and classification based on the acquired sentiment data. It re-evaluates tags based on the sentiment data and makes changes or additions as needed.

[1254] Input: Generated user sentiment data, list of tags assigned to the document.

[1255] Output: Tag lists and categorized documents adjusted based on sentiment data.

[1256] Specific example: If a user is determined to be experiencing stress, documents with tags or categories related to relaxation will be prioritized for display.

[1257] Step 7:

[1258] Tag display by device

[1259] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify or add them as needed.

[1260] Input: A list of tags adjusted based on sentiment data.

[1261] Output: A list of tags displayed on the user's screen.

[1262] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are displayed on the user's screen.

[1263] Step 8:

[1264] Server-based document classification

[1265] The server categorizes documents based on established tags and sentiment data. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[1266] Input: Tag list adjusted based on sentiment data, document file.

[1267] Output: Category information for classified documents.

[1268] Specific example: "Sales Report.pdf" is categorized as "Sales Report," "Fiscal Year 2023," etc.

[1269] Step 9:

[1270] User search and browsing

[1271] The user enters specific keywords to perform a search. The terminal sends the entered keywords to the server, which searches its database for relevant documents. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[1272] Input: Search keywords entered by the user.

[1273] Output: Customized search results list.

[1274] Specific example: When a user searches using the keyword "new customer," relevant documents are displayed preferentially, and "Sales Report.pdf" is found among them.

[1275] (Application Example 2)

[1276] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[1277] In today's information society, a vast amount of documents are stored electronically, making it difficult to efficiently search for necessary documents due to their sheer volume. Furthermore, conventional document management systems rely solely on simple keyword matching without considering the user's emotional state, making it difficult to quickly provide the information the user truly needs. Therefore, there is a need for a system that utilizes user emotional data to customize search results and provide a more personalized search experience.

[1278] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1279] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents using the tags, means for generating user sentiment data, means for adjusting document tags based on sentiment data, means for displaying the generated tags, and means for searching for documents classified based on searched keywords and adjusted tags. This enables automatic document classification and customization of search results based on user sentiment.

[1280] "Means of receiving documents" refers to a function that allows users to upload document data to a server via their device.

[1281] "Means for generating summaries from documents" refers to functions that use natural language processing technology to shorten and concisely display the main content of a document.

[1282] "Methods for extracting important keywords" refer to functions that extract frequently occurring and important information from document summaries.

[1283] "Means of tagging documents" refers to a function that automatically generates and assigns tags related to a document based on extracted keywords.

[1284] "A means of classifying documents using tags" refers to a function that sorts documents into specific categories based on the tags that have been generated.

[1285] "Means for generating user emotion data" refers to a function that analyzes a user's search history and operation history and generates data on their emotional state.

[1286] "Means for adjusting document tags based on sentiment data" refers to a function that dynamically changes the tags and classifications assigned to a document based on generated sentiment data.

[1287] "Means for displaying generated tags" refers to a function that displays a list of tags so that users can review and modify them.

[1288] "Means for searching for documents categorized based on searched keywords and adjusted tags" refers to a function that searches a database for appropriate documents based on keywords and adjusted tags entered by the user and displays the results.

[1289] The system that realizes this invention is designed to allow users to search for documents in an efficient and personalized manner. Specific embodiments are described below.

[1290] Hardware and software to be used

[1291] This system is recommended to use the following hardware and software.

[1292] Hardware:

[1293] CPU: Intel i7 or higher recommended

[1294] Memory: 16GB or more

[1295] Storage: 500GB SSD

[1296] User devices: PCs, smartphones, tablets, etc.

[1297] software:

[1298] OS: Windows 10 or Linux Ubuntu 20.04

[1299] Libraries: SpaCy (en_core_web_sm model), TextBlob, Python 3.x

[1300] System configuration and operation

[1301] 1. Upload document

[1302] Users upload documents and files they want to categorize to the server via their device.

[1303] A dedicated web interface or smartphone app can be used.

[1304] 2. Summary generation

[1305] The server receives the uploaded document and generates a summary using natural language processing technology (such as SpaCy).

[1306] For example, summarization algorithms such as TextRank and BERT are used to shorten and concisely display the main content of a document.

[1307] 3. Keyword Extraction

[1308] The process of extracting key keywords from the summary uses TF-IDF, Word2Vec, and SpaCy.

[1309] For example, keywords such as "sales revenue" and "new customers" can be extracted from the content of uploaded documents.

[1310] 4. Tag generation

[1311] Based on the extracted keywords, generate tags relevant to the document.

[1312] This automatically adds tags such as "increased sales" and "new customers" to documents.

[1313] 5. Generation and adjustment of emotional data

[1314] The server generates sentiment data using a sentiment engine based on the user's operation history and search history.

[1315] For example, the document's tags are adjusted based on whether the user has recently viewed many negative reviews.

[1316] 6. Displaying tags and classifying documents

[1317] The device displays the generated tag list so that the user can review it and make corrections as needed.

[1318] The server categorizes documents into specific categories based on confirmed tags and sentiment data.

[1319] 7. Search and Browsing

[1320] When a user performs a search using a specific keyword, the device sends that keyword to the server.

[1321] The server searches the database for relevant documents and displays customized search results based on sentiment data.

[1322] Specific example

[1323] For example, if a user uploads a "Sales Report for the First Quarter of Fiscal Year 2023," the server receives the document, generates a summary, and displays a summary like the following:

[1324] summary:

[1325] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1326] Furthermore, keywords such as "FY2023," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" are extracted from the summary, and tags are generated based on these. After tag generation, if the user has recently seen many positive reviews, a "Positive" tag is added, adjusting the search results according to the user's sentiment.

[1327] Example of a prompt

[1328] Please write the following program to generate a user sentiment score based on reviews and adjust the document tags accordingly.

[1329] 1. Use Spasi to summarize the document.

[1330] 2. Extract keywords from the summary.

[1331] 3. Tags are generated based on the keywords extracted by Spasi.

[1332] 4. Calculate sentiment data from user reviews. (Examples: 'This product is amazing!', 'Not satisfied', 'Excellent quality')

[1333] 5. Adjust the tags generated based on the calculated sentiment data. (Add a "negative" tag if the sentiment is negative, and a "positive" tag if it's positive.)

[1334] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1335] Step 1:

[1336] The server receives documents uploaded by users via their terminals. The server receives the document data (PDF, DOCX, TXT, etc.) as input and stores it within the document management system. At the same time, the document's metadata (title, creation date, etc.) is also saved.

[1337] Step 2:

[1338] The server analyzes the received document data using natural language processing techniques (such as SpaCy) and generates a document summary. It analyzes the document text as input, extracts the main content, and generates a summary. It generates a summary text as output and passes it on to the next process. Specifically, this involves sentence segmentation, importance calculation, and selection of summary sentences.

[1339] Step 3:

[1340] The server extracts key keywords from the generated summary. Based on the summary text as input, it selects key keywords using algorithms such as TF-IDF and Word2Vec. It then generates a list of key keywords as output. Specifically, this involves calculating word frequency, applying inverse document frequency, and calculating weights.

[1341] Step 4:

[1342] The server automatically generates and assigns tags related to the document based on the extracted keywords. It uses a keyword list as input to select tags to be assigned to the document. It generates a tag list as output and associates it with the document. Specifically, this involves keyword mapping and the generation of category tags.

[1343] Step 5:

[1344] The server generates sentiment data based on the user's operation history and search history. It analyzes the user history data as input and calculates a sentiment score using a sentiment engine. It generates sentiment data as output and incorporates it into the next process. Specifically, this involves analyzing the user's behavior log and applying the sentiment scoring model.

[1345] Step 6:

[1346] The server adjusts the document's tags based on the generated sentiment data. It optimizes the tags using the tag list and sentiment data as input. It generates an adjusted tag list as output and re-assigns the tags to the document. Specifically, this involves adjusting the weighting of tags and modifying the tagging based on sentiment.

[1347] Step 7:

[1348] The terminal displays the generated tag list so that the user can review it. It displays the adjusted tag list as input and allows the user to provide feedback on the tags. It receives user feedback data as output. Specifically, it generates the tag list display interface and provides user input fields.

[1349] Step 8:

[1350] The server categorizes documents based on confirmed tags and sentiment data. It uses the final tag list and sentiment data as input to categorize documents. The categorization information is saved to the document database as output. Specifically, the process involves mapping tags to sentiment data and applying a categorization algorithm.

[1351] Step 9:

[1352] When a user performs a search using a specific keyword, the device sends the entered keyword to the server. Upon receiving the search keyword as input, the server searches its database for relevant documents. It then generates a list of search results and sends it back to the device. Specific operations include keyword matching, ranking of search results, and optimization of results based on sentiment data.

[1353] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1354] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1355] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[1356] [Fourth Embodiment]

[1357] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[1358] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1359] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1360] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[1361] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[1362] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[1363] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[1364] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[1365] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[1366] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1367] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1368] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[1369] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1370] This invention relates to a system for automatically sorting and efficiently searching documents. This invention enables consistent document management that is not influenced by human subjectivity, and significantly improves search efficiency.

[1371] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[1372] Specific processing of the system

[1373] 1. User uploads documents

[1374] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[1375] 2. Server-driven summary generation

[1376] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, it uses summarization algorithms such as TextRank and BERT to shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[1377] 3. Keyword extraction by the server

[1378] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[1379] 4. Tag generation by the server

[1380] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[1381] 5. Displaying tags on the device

[1382] The device displays a list of generated tags so that the user can review them. The user can then review the displayed tags and make any necessary corrections.

[1383] 6. Server-based document classification

[1384] The server categorizes documents based on the identified tags. Specifically, documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information." This information is stored in a database and used for later searches.

[1385] 7. User search and browsing

[1386] The user performs a search using specific keywords. The terminal sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the terminal, allowing the user to quickly find and view the documents they need.

[1387] Specific example

[1388] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[1389] summary:

[1390] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1391] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[1392] Users can review and modify these tags. Ultimately, the server categorizes documents based on these tags and stores them in the database. When a user searches using keywords such as "new customer," the server can quickly find relevant documents and display the results.

[1393] This allows for efficient document searching and management, without relying on human subjectivity.

[1394] The following describes the processing flow.

[1395] Step 1:

[1396] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[1397] Step 2:

[1398] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[1399] Step 3:

[1400] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[1401] Step 4:

[1402] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[1403] Step 5:

[1404] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[1405] Step 6:

[1406] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[1407] Step 7:

[1408] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[1409] Step 8:

[1410] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[1411] Step 9:

[1412] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[1413] Step 10:

[1414] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[1415] Step 11:

[1416] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[1417] Step 12:

[1418] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[1419] Step 13:

[1420] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[1421] Step 14:

[1422] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[1423] Step 15:

[1424] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[1425] Step 16:

[1426] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[1427] (Example 1)

[1428] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1429] Currently, many companies and individuals face the challenge of managing vast amounts of documents. In this situation, document classification and searching are often done manually, which is time-consuming and labor-intensive. Furthermore, management based on human subjectivity can lead to problems with the consistency of classification and the accuracy of searches. Therefore, there is a growing need for systems that can automatically organize documents and efficiently search them.

[1430] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1431] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for displaying the tags on a terminal, means for classifying documents based on the tags, and means for searching for documents classified based on the searched keywords. This enables automatic organization of documents, consistent document management, and efficient searching.

[1432] "Means of receiving documents" refers to the interface and communication means for users to upload document files to the server.

[1433] A "means for generating summaries from documents" refers to a summarization algorithm that uses natural language processing technology to summarize the content of a document, extract only the important information, and condense it into short sentences.

[1434] "Methods for extracting important keywords from summaries" refer to algorithms and processes for selecting frequently occurring words and semantically important words from generated summaries.

[1435] "Means for tagging documents based on extracted keywords" refers to algorithms and systems that automatically generate relevant tags based on extracted keywords.

[1436] "Means of displaying tags on a device" refers to a display screen and interface that allows users to check and modify the generated tags.

[1437] "Means for classifying documents based on tags" refers to algorithms and database management systems that automatically classify documents into appropriate categories using established tags.

[1438] "Means for searching for documents classified based on searched keywords" refers to a system and process that searches a database for relevant documents based on keywords entered by the user and displays the results.

[1439] This invention relates to a system for automatically sorting and efficiently searching documents. To implement this system, the following specific hardware and software are used to perform each processing step.

[1440] First, users utilize a dedicated web interface or application (e.g., a web browser) on a device such as a personal computer or smartphone. Through this, users can upload document files they wish to sort to the server. Specifically, using a common web browser such as Google Chrome or Mozilla Firefox, users select the documents they want to upload from a file selection dialog and click the "Upload" button.

[1441] The server receives document files sent by users and generates summaries of those documents using natural language processing (NLP) techniques. This process applies summarization algorithms such as TextRank and BERT. The server uses these algorithms to extract and analyze the key content of the document and create a concise summary.

[1442] Next, the server extracts key keywords from the generated summary. This process utilizes algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. This selects frequently occurring or semantically important words from the summary as keywords.

[1443] The server generates tags related to the document based on the extracted keywords. For example, if a sales report is uploaded, the server generates a summary such as "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this, it extracts keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies." Then, it generates tags such as "fiscal year 2023," "sales increase," and "new customers" from these keywords and attaches them to the document.

[1444] The device displays the generated tag list to the user. Through this display, the user can review the tags and modify or add them as needed. The display interface is intuitive and user-friendly, designed for easy operation.

[1445] The server then categorizes the documents based on the identified tags. This process utilizes a tag classification algorithm and a database management system. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[1446] Ultimately, when a user searches for a document using specific keywords, the terminal sends the entered keywords to the server. The server searches the database for the relevant documents and displays the search results on the terminal. This allows the user to quickly find and view the documents they need.

[1447] An example of a specific prompt message is as follows:

[1448] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1449] Please extract the key keywords from the summary above.

[1450] Please generate tags appropriate for the document in question.

[1451] Based on this, please categorize the document as a "sales report".

[1452] As described above, this system enables automatic document sorting and efficient searching. This allows for consistent document management and significantly improves search efficiency.

[1453] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1454] Step 1:

[1455] User upload of documents

[1456] Users upload document files using a dedicated web interface or application via a device such as a personal computer or smartphone. The user launches a browser, selects the document file from the file selection dialog, and clicks the "Upload" button.

[1457] Input: Document file selected by the user

[1458] Output: Document file sent to the server

[1459] Step 2:

[1460] Receiving documents from the server

[1461] The server receives document files sent by users. The received files are temporarily stored for subsequent processing.

[1462] Input: Uploaded document file

[1463] Output: Saved document file

[1464] Step 3:

[1465] Server-based summary generation

[1466] The server analyzes stored document files and generates summaries using natural language processing techniques. Specific algorithms used include TextRank and BERT. The server analyzes the document content, extracts key information, and generates a concise summary.

[1467] Input: Saved document file

[1468] Data processing: Summarization using natural language processing techniques (TextRank and BERT)

[1469] Output: Generated summary

[1470] Step 4:

[1471] Keyword extraction by the server

[1472] The server extracts key keywords from the generated summary. This process uses tools such as TF-IDF, Word2Vec, and SpaCy. It calculates the frequency and importance of words included in the summary and selects keywords.

[1473] Input: Generated summary

[1474] Data processing: Keyword extraction using TF-IDF, Word2Vec, and SpaCy.

[1475] Output: Extracted keyword list

[1476] Step 5:

[1477] Server-based tag generation

[1478] The server generates relevant tags based on the extracted keywords. It analyzes the keywords and creates tags based on specific conditions.

[1479] Input: Extracted keyword list

[1480] Data processing: Keyword-based tag generation algorithm

[1481] Output: Generated tag list

[1482] Step 6:

[1483] Tag display by device

[1484] The terminal displays a list of tags sent from the server to the user. The user reviews the displayed tags and makes corrections or additions as needed.

[1485] Input: Tag list sent from the server

[1486] Output: Tag list displayed to the user

[1487] Step 7:

[1488] Server-based document classification

[1489] The server classifies documents into specific categories based on confirmed tags. It uses a tag classification algorithm to assign documents to the appropriate categories and saves them to the database.

[1490] Input: Confirmed tag list

[1491] Data processing: Tag-based document classification algorithm

[1492] Output: Category information of documents stored in the database

[1493] Step 8:

[1494] User search and browsing

[1495] The user enters a specific keyword into the search bar on their device to search for documents. The device sends the entered keyword to the server, which searches its database for the relevant documents. The search results are displayed on the device, allowing the user to view the necessary documents.

[1496] Input: Search keywords entered by the user

[1497] Output: List of documents displayed as search results

[1498] Specific operation: User enters keywords into the search bar → Terminal sends keywords to the server → Server searches for documents → Search results are displayed on the terminal.

[1499] (Application Example 1)

[1500] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1501] Traditional document management systems suffer from the drawback of requiring many manual and time-consuming operations for document uploading, summary generation, keyword extraction, tagging, classification, and searching. Furthermore, for use in field settings such as logistics centers, efficient and intuitive operation is required, necessitating a flexible system utilizing smart devices. Smart search methods, such as voice commands, are also in demand.

[1502] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1503] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents, means for displaying tags, means for searching for classified documents based on searched keywords, means for reading documents via a smart device and uploading them to the cloud, and means for displaying search results on the smart device's display. This enables the automation of the document management process and efficient and intuitive document management using smart devices.

[1504] "Means of receiving documents" refers to the functions or devices that allow users to upload documents.

[1505] "Means for generating summaries from documents" refers to functions or devices that shorten and concisely express the main content of long documents.

[1506] "Means for extracting important keywords from a summary" refers to a function or device that selects important words from a generated summary.

[1507] "Means of tagging documents based on extracted keywords" refers to functions or devices that automatically assign labels related to documents based on keywords.

[1508] "Means of classifying documents" refers to functions or devices that categorize documents into specific categories based on their content.

[1509] "Means of displaying tags" refers to functions or devices that visually present tags assigned to a document to the user.

[1510] "Means for searching for documents classified based on searched keywords" refers to functions or devices that find appropriate documents based on keywords entered by the user.

[1511] "Means of reading documents via smart devices and uploading them to the cloud" refers to functions or devices that use smart glasses or other devices to scan documents and send them to a cloud server via the internet.

[1512] "Means of displaying search results on a smart device's display" refers to functions or devices that visually present search results to the user on the screen of a device such as smart glasses.

[1513] To implement this invention, the following system configuration and specific processing are used. The embodiment combines a server, a smart device (e.g., smart glasses), and cloud storage.

[1514] The server forms the core of a system that automatically sorts and efficiently searches documents, and has the following functions:

[1515] 1. Means of receiving documents:

[1516] Users use the smart glasses' camera to photograph documents, and the captured document data is uploaded to the cloud via the internet. The application on the smart glasses provides an interface for easily capturing documents and sending them to the server.

[1517] 2. Means for generating summaries from documents:

[1518] The server processes the received document and generates a summary using natural language processing techniques (e.g., TextRank or BERT). This summarization process shortens and concisely displays the main points of the document. For example, if a logistics instruction sheet is uploaded, a summary such as "Inventory list for Q2 2023 increased by 10% year-on-year" might be generated.

[1519] 3. Methods for extracting key keywords from summaries:

[1520] Next, the server extracts key keywords from the generated summary. This involves using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy) to identify and extract important concepts within the document.

[1521] 4. A means of tagging documents based on extracted keywords:

[1522] The system automatically generates relevant tags from the extracted keywords and adds them to the document. For example, tags such as "FY2023," "Q2," and "Inventory Increase" will be generated.

[1523] 5. Means of classifying documents:

[1524] The server categorizes documents based on the identified tags. This assigns documents to categories such as "periodical reports" and "inventory management." This information is stored in a database and used for later searches.

[1525] 6. Means of displaying tags:

[1526] The tag list generated by the server is displayed on the smart glasses' screen. The user can view the displayed tags and make corrections as needed.

[1527] 7. Means for searching for documents categorized based on searched keywords:

[1528] Users can use the voice command function of their smart glasses to perform searches using specific keywords. The smart glasses' display shows tags and summaries of relevant documents, allowing users to quickly find the documents they are looking for.

[1529] Hardware and software to use

[1530] Hardware: Smart glasses (e.g., Google Glass)

[1531] software:

[1532] Natural language processing models: spaCy, TextRank, BERT

[1533] Cloud storage: AWS S3, Google Cloud Storage

[1534] Specific example

[1535] When logistics center staff receive a new delivery order, they use smart glasses to photograph the document and upload it. The cloud server summarizes the document and generates tags. When they say the voice command "Search for new delivery orders," the smart glasses screen displays a list of relevant documents.

[1536] Example of a prompt

[1537] "Shipping Instructions: Generate a summary inventory list for Q2 2023 and extract tags. Output: Inventory List Summary: Inventory volume increased by 10% year-on-year in Q2 2023. Related tags: 2023, Q2, Inventory volume, 10% increase"

[1538] In this way, the document management process can be automated and streamlined, supporting the operations of the logistics center. This improves work efficiency and allows users to quickly find the documents they need.

[1539] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1540] Step 1:

[1541] The user uses smart glasses to photograph a document. This document data becomes the input. The user uses the smart glasses application to easily capture the document, and the captured document is uploaded to the cloud via the internet. The output of this process is the document data stored in cloud storage.

[1542] Step 2:

[1543] The server retrieves document data stored in cloud storage. This document data becomes the input, and the server uses natural language processing techniques (e.g., TextRank, BERT) to generate a summary of the document. The generated summary becomes the output. Specifically, it extracts the main content from the document and summarizes it in a shortened form.

[1544] Step 3:

[1545] The server extracts key keywords from the generated summary. This summary data becomes the input, and important concepts within the document are identified using TF-IDF, Word2Vec, and natural language processing libraries (e.g., spaCy). The extracted keywords become the output. Specifically, the frequency and context of the keywords are analyzed, and their importance is evaluated.

[1546] Step 4:

[1547] The server assigns tags to documents based on extracted keywords. This keyword data serves as input, and appropriate tags are automatically generated. The generated tag list is the output. Specifically, it searches the database for tags highly relevant to the keywords and assigns them to the documents.

[1548] Step 5:

[1549] The server categorizes documents. This tag list serves as input, and the server assigns documents to categories such as "periodical reports" and "inventory management." The category information of the classified documents is output. Specifically, it maps documents to the appropriate category based on their tags.

[1550] Step 6:

[1551] The server sends the generated tag list to the terminal, which displays it on the smart glasses' screen. This tag list serves as input, and the tags are displayed so that the user can see them. The displayed tag list becomes the output. Specifically, the tags are displayed on the smart glasses' HUD (Heads-Up Display).

[1552] Step 7:

[1553] The user performs a search using a specific keyword via voice command. This keyword becomes the input, and the server searches the database for relevant documents. The searched document list is the output. Specifically, the system quickly finds documents that match or are related to the search keyword from the database and displays the results on the smart glasses' display.

[1554] Through the above processes, users can efficiently manage documents using smart devices. Furthermore, the search function allows them to quickly find the documents they need.

[1555] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1556] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it is an invention that combines a sentiment engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results.

[1557] To explain the implementation of the invention, a program for this system is generated, and the processing of that program is described in natural language.

[1558] Specific processing of the system

[1559] 1. User uploads documents

[1560] Users upload documents and files they want to sort to the server via their device. They can easily select and send files using a dedicated web interface or application.

[1561] 2. Server-driven summary generation

[1562] The server receives uploaded documents and uses natural language processing (NLP) techniques to create summaries. Specifically, summarization algorithms such as TextRank and BERT are used to generate summaries that shorten and concisely display the main points of the document. For example, if a sales report is uploaded, the server will generate a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[1563] 3. Keyword extraction by the server

[1564] The server extracts key keywords from the generated summary. This process uses algorithms and filtering techniques such as TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, and SpaCy. For example, keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" are extracted.

[1565] 4. Tag generation by the server

[1566] The server generates tags related to the document based on the extracted keywords. For example, tags such as "FY2023," "Increased Sales," and "New Customers" are automatically generated and attached to the document.

[1567] 5. Acquisition of sentiment data by the server

[1568] The server is equipped with an emotion engine that generates emotion data from user input and operation history. For example, the emotion engine analyzes the user's emotion data based on information such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency.

[1569] 6. Server-side adjustment of tags and sentiment data

[1570] The server uses sentiment data generated by the sentiment engine to adjust document tagging and classification. For example, if a user is determined to be stressed based on their recent search history, the server will prioritize displaying documents related to relaxation.

[1571] 7. Displaying tags on the device

[1572] The device displays a list of generated tags so that the user can review them. The user can review the displayed tags and modify them as needed.

[1573] 8. Server-based document classification

[1574] The server categorizes documents based on established tags and sentiment data. Specifically, documents are assigned to categories such as "Sales Reports," "Annual," and "Customer Information." This information is stored in a database and used for later searches.

[1575] 9. User search and browsing

[1576] The user performs a search using specific keywords. The device sends the entered keywords to the server, which searches the database for relevant documents. The search results are displayed on the device, allowing the user to quickly find and view the documents they need. During this process, an emotion engine customizes the search results based on the user's emotions, prioritizing documents that meet the user's needs.

[1577] Specific example

[1578] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary like the following:

[1579] summary:

[1580] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1581] Next, the server extracts keywords such as "FY2023," "Q1," "Sales," "20% increase," "New customers," and "15 companies" from this summary. Then, it generates tags such as "FY2023," "Sales increase," and "New customers" from these keywords and adds them to the document.

[1582] The server is equipped with an emotion engine that generates emotion data from the user's operation history and search history. Based on this emotion data, tagging and document classification are adjusted.

[1583] Ultimately, when a user performs a search using keywords such as "new customer," the server can quickly search for relevant documents, taking sentiment data into consideration, and display the results. In this way, it becomes possible to efficiently search and manage documents without relying on human subjectivity.

[1584] The following describes the processing flow.

[1585] Step 1:

[1586] Users select documents from a web interface or dedicated application and upload them to the server. Users select local files through a file selection dialog and click the "Upload" button.

[1587] Step 2:

[1588] The device sends the selected file to the server. The file is uploaded and transferred to the server via an HTTP request.

[1589] Step 3:

[1590] The server saves the received documents to storage. The saved documents are stored in a temporary directory and used for subsequent processing.

[1591] Step 4:

[1592] The server passes the stored documents to a natural language processing (NLP) library to generate a summary. Specifically, summarization algorithms such as TextRank and BERT are used to generate a concise summary that expresses the main content of the document.

[1593] Step 5:

[1594] The server saves the generated summary to the database. The summary is stored in the database along with related information because it will be used in subsequent processing.

[1595] Step 6:

[1596] The server reviews the summary and extracts key keywords from it. Algorithms such as TF-IDF, Word2Vec, and SpaCy are used to extract important words and phrases from the summary.

[1597] Step 7:

[1598] The server saves the extracted keywords to a database. The extracted keywords are recorded in the database because they are used as the basis for tag generation.

[1599] Step 8:

[1600] The server generates tags related to the document based on the extracted keywords. For example, based on keywords such as "2023 fiscal year" and "new customer," tags such as "by fiscal year" and "customer information" are automatically generated.

[1601] Step 9:

[1602] The server saves the generated tags to the database. Since tags are used for document classification and searching, they are stored in the database along with related documents.

[1603] Step 10:

[1604] The device displays the generated tags to the user. The user can review the displayed tags through the web interface and modify them as needed.

[1605] Step 11:

[1606] The server uses an emotion engine to generate emotion data from the user's operation history and search history. The emotion engine determines the user's emotional state and stores that data.

[1607] Step 12:

[1608] The server adjusts tagging and document classification based on emotional data. For example, if a user is feeling stressed, it will prioritize displaying relaxation-related documents.

[1609] Step 13:

[1610] The server classifies documents into appropriate categories based on the identified tags. Based on the tag information, documents are assigned to categories such as "Sales Reports" or "Customer Information."

[1611] Step 14:

[1612] The server saves the classification results to the database. The classified document information is stored in the database to enable rapid searching.

[1613] Step 15:

[1614] The user enters keywords into the search bar and performs a search. The entered keywords are sent to the server via the device.

[1615] Step 16:

[1616] The server searches the database for the relevant documents. The search engine matches the documents based on keywords and extracts the relevant documents.

[1617] Step 17:

[1618] The server sends the search results to the terminal. The list of documents obtained as search results is displayed to the user.

[1619] Step 18:

[1620] The user selects the document they need and views its details. The user can then click on the relevant document from the displayed document list to view its details.

[1621] Step 19:

[1622] The emotion engine customizes search results based on the user's emotional data. For example, if the emotion engine determines that the user is stressed, relaxation-related documents will be displayed preferentially.

[1623] (Example 2)

[1624] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1625] Conventional document management systems statically tag and classify documents, failing to reflect user emotions and intentions. As a result, it is difficult for users to quickly and accurately search for the documents they are looking for. Furthermore, there is a lack of methods to customize document search results and improve the user experience by utilizing sentiment data. This invention aims to achieve more efficient and customized document management and retrieval by dynamically adjusting document tagging and classification using user sentiment data.

[1626] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1627] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for acquiring user sentiment data, means for adjusting the classification of tags and documents based on the acquired sentiment data, means for classifying documents, means for displaying tags, and means for searching for classified documents based on searched keywords. This enables dynamic tagging and document classification based on user sentiment data.

[1628] "Means of receiving documents" refers to a system that sends documents uploaded by users to a server and has the function of receiving those documents on the server.

[1629] "Means for generating summaries from documents" refers to a function in which a server analyzes the entire content of a document and automatically generates a concise summary that highlights the main points.

[1630] "Method for extracting important keywords from a summary" refers to a function where the server analyzes the generated summary and extracts important words and phrases that represent the content of the document.

[1631] "A means of tagging documents based on extracted keywords" refers to a system that analyzes keywords and automatically assigns highly relevant tags to documents based on those keywords.

[1632] "Means for acquiring user sentiment data" refers to a function that analyzes the user's operation history and search history to collect data for inferring the user's emotional state.

[1633] "Means for adjusting tags and document classifications based on acquired sentiment data" refers to a system that dynamically changes document tags and classifications using collected sentiment data, performing appropriate tagging and document classification according to the user's emotional state.

[1634] "Means of classifying documents" refers to the function of dividing documents into specific categories or folders based on tags assigned to them.

[1635] "Means for displaying tags" refers to a function that visually displays tags generated by the server on the user interface, allowing users to review and modify them.

[1636] "Means for searching for documents classified based on searched keywords" refers to a function in which the server searches the database for relevant documents based on the search keywords entered by the user and displays the results.

[1637] This invention relates to a system for automatically sorting and efficiently searching documents. In particular, it incorporates an emotion engine that uses user sentiment data to adjust document tagging and classification, thereby customizing search results. The specific implementation method of this system is described below.

[1638] Hardware and software used

[1639] The server is the entity responsible for storing, analyzing, tagging, sentiment analysis, and retrieving document data. Typical software includes libraries and frameworks for natural language processing (NLP) (e.g., TextRank, BERT, SpaCy), algorithms for keyword extraction such as TF-IDF and Word2Vec, and sentiment analysis engines.

[1640] A terminal is a device used by users to upload documents, check tags, and receive search results. A common web browser or a dedicated application is used as the interface.

[1641] Users are those who utilize the system and perform operations such as uploading documents, checking and modifying tags, and searching.

[1642] Process Overview

[1643] Upload document

[1644] Users upload documents and files they want to sort to the server via their device. They can easily select files using a dedicated web interface or application and send them to the server by clicking the upload button.

[1645] Summary generation

[1646] The server analyzes the received document and generates a summary using natural language processing techniques such as TextRank and BERT. This summary is used to concisely grasp the main points of the document.

[1647] Keyword extraction

[1648] The server extracts important keywords from the generated summary using algorithms such as TF-IDF, Word2Vec, and SpaCy. This extracts words and phrases that represent the content of the document.

[1649] Tag generation

[1650] The server generates tags related to the document based on the extracted keywords. Using a tag generation algorithm, it converts the extracted keywords into tags and assigns them to the document.

[1651] Acquisition of emotional data

[1652] The server is equipped with an emotion engine that acquires emotion data based on the user's operation history and search history. The emotion engine analyzes the user's emotions from data such as what kind of searches the user performed and which documents they viewed.

[1653] Adjusting tags and document classification

[1654] The server adjusts document tagging and classification based on acquired emotional data. If the server determines that the user is stressed, it prioritizes displaying documents with tags and classifications related to relaxation.

[1655] Tag display

[1656] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify them as needed.

[1657] Document classification

[1658] The server categorizes documents based on established tags and sentiment data. Documents are categorized into categories such as "Sales Reports," "Annual," and "Customer Information," and stored in the database.

[1659] Search and browsing

[1660] The user enters specific keywords to perform a search. The server searches the database for relevant documents and sends the results to the user's terminal. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[1661] Specific example

[1662] For example, consider a case where a user uploads a "sales report" to the server. The server receives the document, generates a summary, and obtains a summary such as, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired." From this summary, keywords such as "fiscal year 2023," "first quarter," "sales," "20% increase," "new customers," and "15 companies" are extracted, and tags such as "fiscal year 2023," "sales increase," and "new customers" are generated.

[1663] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is seeking relaxation and will prioritize displaying related documents.

[1664] Examples of prompts for generative AI models

[1665] "After uploading the Sales Report.pdf file, perform text summarization, keyword extraction, and tag generation, then adjust the tags and document classification using sentiment data."

[1666] Thus, the system of the present invention can significantly improve the efficiency of document management and retrieval by utilizing user sentiment data.

[1667] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1668] Step 1:

[1669] User upload of documents

[1670] Users upload documents and files they want to categorize to the server using their device. Users open a dedicated web interface or application, click the file selection button, and select files such as "Sales Report.pdf". The selected files are sent from the device to the server by clicking the upload button.

[1671] Input: A document file uploaded by the user.

[1672] Output: Document file sent to the server.

[1673] Step 2:

[1674] Server-based summary generation

[1675] The server receives the uploaded document and determines the file format (PDF, Word, text, etc.). It extracts the document content using an appropriate parser and generates a document summary using natural language processing techniques (TextRank or BERT). The server analyzes the document content and creates a concise summary that includes the main points.

[1676] Input: Document file sent to the server.

[1677] Output: A summary of the document.

[1678] Specific example: The server analyzes "Sales Report.pdf" and generates a summary stating, "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. 15 new customers were acquired."

[1679] Step 3:

[1680] Keyword extraction by the server

[1681] The server extracts key keywords from the generated summary. It uses algorithms such as TF-IDF, Word2Vec, and SpaCy to analyze the summary and select frequently occurring and important words.

[1682] Input: A summary of the document.

[1683] Output: A list of extracted important keywords.

[1684] Specific example: Extract keywords such as "FY2023," "First Quarter," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" from the summary.

[1685] Step 4:

[1686] Server-based tag generation

[1687] The server generates tags related to the document based on the extracted keywords. It converts the keywords into the appropriate tag format and assigns them to the document. A tag generation algorithm is used to automatically create relevant tags from the extracted keywords.

[1688] Input: A list of important keywords extracted.

[1689] Output: A list of tags assigned to the document.

[1690] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are generated.

[1691] Step 5:

[1692] Acquisition of emotional data by a server

[1693] The server is equipped with an emotion engine that generates emotion data based on the user's operation history and search history. It analyzes data such as what kind of searches the user performed, which documents they viewed, the time of day, and frequency, in order to infer the user's emotional state.

[1694] Input: Data such as user activity history, search history, referenced documents, time of day, and frequency.

[1695] Output: Generated user sentiment data.

[1696] Specific example: If a user frequently searches using the keyword "stress management," the emotion engine will determine that the user is experiencing stress.

[1697] Step 6:

[1698] Server-based adjustment of tag and document classification

[1699] The server adjusts document tagging and classification based on the acquired sentiment data. It re-evaluates tags based on the sentiment data and makes changes or additions as needed.

[1700] Input: Generated user sentiment data, list of tags assigned to the document.

[1701] Output: Tag lists and categorized documents adjusted based on sentiment data.

[1702] Specific example: If a user is determined to be experiencing stress, documents with tags or categories related to relaxation will be prioritized for display.

[1703] Step 7:

[1704] Tag display by device

[1705] The device displays a list of tags generated through the user interface to the user. The user can review the displayed tags and modify or add them as needed.

[1706] Input: A list of tags adjusted based on sentiment data.

[1707] Output: A list of tags displayed on the user's screen.

[1708] Specific example: Tags such as "Fiscal Year 2023," "Increased Sales," and "New Customers" are displayed on the user's screen.

[1709] Step 8:

[1710] Server-based document classification

[1711] The server categorizes documents based on established tags and sentiment data. Documents are assigned to categories such as "Sales Reports," "Annual Records," and "Customer Information," and stored in the database.

[1712] Input: Tag list adjusted based on sentiment data, document file.

[1713] Output: Category information for classified documents.

[1714] Specific example: "Sales Report.pdf" is categorized as "Sales Report," "Fiscal Year 2023," etc.

[1715] Step 9:

[1716] User search and browsing

[1717] The user enters specific keywords to perform a search. The terminal sends the entered keywords to the server, which searches its database for relevant documents. The search results are customized based on the user's sentiment data, allowing the user to quickly find and view the documents they need.

[1718] Input: Search keywords entered by the user.

[1719] Output: Customized search results list.

[1720] Specific example: When a user searches using the keyword "new customer," relevant documents are displayed preferentially, and "Sales Report.pdf" is found among them.

[1721] (Application Example 2)

[1722] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1723] In today's information society, a vast amount of documents are stored electronically, making it difficult to efficiently search for necessary documents due to their sheer volume. Furthermore, conventional document management systems rely solely on simple keyword matching without considering the user's emotional state, making it difficult to quickly provide the information the user truly needs. Therefore, there is a need for a system that utilizes user emotional data to customize search results and provide a more personalized search experience.

[1724] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[1725] In this invention, the server includes means for receiving documents, means for generating summaries from documents, means for extracting important keywords from summaries, means for tagging documents based on the extracted keywords, means for classifying documents using the tags, means for generating user sentiment data, means for adjusting document tags based on sentiment data, means for displaying the generated tags, and means for searching for documents classified based on searched keywords and adjusted tags. This enables automatic document classification and customization of search results based on user sentiment.

[1726] "Means of receiving documents" refers to a function that allows users to upload document data to a server via their device.

[1727] "Means for generating summaries from documents" refers to functions that use natural language processing technology to shorten and concisely display the main content of a document.

[1728] "Methods for extracting important keywords" refer to functions that extract frequently occurring and important information from document summaries.

[1729] "Means of tagging documents" refers to a function that automatically generates and assigns tags related to a document based on extracted keywords.

[1730] "A means of classifying documents using tags" refers to a function that sorts documents into specific categories based on the tags that have been generated.

[1731] "Means for generating user emotion data" refers to a function that analyzes a user's search history and operation history and generates data on their emotional state.

[1732] "Means for adjusting document tags based on sentiment data" refers to a function that dynamically changes the tags and classifications assigned to a document based on generated sentiment data.

[1733] "Means for displaying generated tags" refers to a function that displays a list of tags so that users can review and modify them.

[1734] "Means for searching for documents categorized based on searched keywords and adjusted tags" refers to a function that searches a database for appropriate documents based on keywords and adjusted tags entered by the user and displays the results.

[1735] The system that realizes this invention is designed to allow users to search for documents in an efficient and personalized manner. Specific embodiments are described below.

[1736] Hardware and software to be used

[1737] This system is recommended to use the following hardware and software.

[1738] Hardware:

[1739] CPU: Intel i7 or higher recommended

[1740] Memory: 16GB or more

[1741] Storage: 500GB SSD

[1742] User devices: PCs, smartphones, tablets, etc.

[1743] software:

[1744] OS: Windows 10 or Linux Ubuntu 20.04

[1745] Libraries: SpaCy (en_core_web_sm model), TextBlob, Python 3.x

[1746] System configuration and operation

[1747] 1. Upload document

[1748] Users upload documents and files they want to categorize to the server via their device.

[1749] A dedicated web interface or smartphone app can be used.

[1750] 2. Summary generation

[1751] The server receives the uploaded document and generates a summary using natural language processing technology (such as SpaCy).

[1752] For example, summarization algorithms such as TextRank and BERT are used to shorten and concisely display the main content of a document.

[1753] 3. Keyword Extraction

[1754] The process of extracting key keywords from the summary uses TF-IDF, Word2Vec, and SpaCy.

[1755] For example, keywords such as "sales revenue" and "new customers" can be extracted from the content of uploaded documents.

[1756] 4. Tag generation

[1757] Based on the extracted keywords, generate tags relevant to the document.

[1758] This automatically adds tags such as "increased sales" and "new customers" to documents.

[1759] 5. Generation and adjustment of emotional data

[1760] The server generates sentiment data using a sentiment engine based on the user's operation history and search history.

[1761] For example, the document's tags are adjusted based on whether the user has recently viewed many negative reviews.

[1762] 6. Displaying tags and classifying documents

[1763] The device displays the generated tag list so that the user can review it and make corrections as needed.

[1764] The server categorizes documents into specific categories based on confirmed tags and sentiment data.

[1765] 7. Search and Browsing

[1766] When a user performs a search using a specific keyword, the device sends that keyword to the server.

[1767] The server searches the database for relevant documents and displays customized search results based on sentiment data.

[1768] Specific example

[1769] For example, if a user uploads a "Sales Report for the First Quarter of Fiscal Year 2023," the server receives the document, generates a summary, and displays a summary like the following:

[1770] summary:

[1771] "Sales for the first quarter of fiscal year 2023 increased by 20% year-on-year. We acquired 15 new customers."

[1772] Furthermore, keywords such as "FY2023," "Sales Revenue," "20% Increase," "New Customers," and "15 Companies" are extracted from the summary, and tags are generated based on these. After tag generation, if the user has recently seen many positive reviews, a "Positive" tag is added, adjusting the search results according to the user's sentiment.

[1773] Example of a prompt

[1774] Please write the following program to generate a user sentiment score based on reviews and adjust the document tags accordingly.

[1775] 1. Use Spasi to summarize the document.

[1776] 2. Extract keywords from the summary.

[1777] 3. Tags are generated based on the keywords extracted by Spasi.

[1778] 4. Calculate sentiment data from user reviews. (Examples: 'This product is amazing!', 'Not satisfied', 'Excellent quality')

[1779] 5. Adjust the tags generated based on the calculated sentiment data. (Add a "negative" tag if the sentiment is negative, and a "positive" tag if it's positive.)

[1780] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1781] Step 1:

[1782] The server receives documents uploaded by users via their terminals. The server receives the document data (PDF, DOCX, TXT, etc.) as input and stores it within the document management system. At the same time, the document's metadata (title, creation date, etc.) is also saved.

[1783] Step 2:

[1784] The server analyzes the received document data using natural language processing techniques (such as SpaCy) and generates a document summary. It analyzes the document text as input, extracts the main content, and generates a summary. It generates a summary text as output and passes it on to the next process. Specifically, this involves sentence segmentation, importance calculation, and selection of summary sentences.

[1785] Step 3:

[1786] The server extracts key keywords from the generated summary. Based on the summary text as input, it selects key keywords using algorithms such as TF-IDF and Word2Vec. It then generates a list of key keywords as output. Specifically, this involves calculating word frequency, applying inverse document frequency, and calculating weights.

[1787] Step 4:

[1788] The server automatically generates and assigns tags related to the document based on the extracted keywords. It uses a keyword list as input to select tags to be assigned to the document. It generates a tag list as output and associates it with the document. Specifically, this involves keyword mapping and the generation of category tags.

[1789] Step 5:

[1790] The server generates sentiment data based on the user's operation history and search history. It analyzes the user history data as input and calculates a sentiment score using a sentiment engine. It generates sentiment data as output and incorporates it into the next process. Specifically, this involves analyzing the user's behavior log and applying the sentiment scoring model.

[1791] Step 6:

[1792] The server adjusts the document's tags based on the generated sentiment data. It optimizes the tags using the tag list and sentiment data as input. It generates an adjusted tag list as output and re-assigns the tags to the document. Specifically, this involves adjusting the weighting of tags and modifying the tagging based on sentiment.

[1793] Step 7:

[1794] The terminal displays the generated tag list so that the user can review it. It displays the adjusted tag list as input and allows the user to provide feedback on the tags. It receives user feedback data as output. Specifically, it generates the tag list display interface and provides user input fields.

[1795] Step 8:

[1796] The server categorizes documents based on confirmed tags and sentiment data. It uses the final tag list and sentiment data as input to categorize documents. The categorization information is saved to the document database as output. Specifically, the process involves mapping tags to sentiment data and applying a categorization algorithm.

[1797] Step 9:

[1798] When a user performs a search using a specific keyword, the device sends the entered keyword to the server. Upon receiving the search keyword as input, the server searches its database for relevant documents. It then generates a list of search results and sends it back to the device. Specific operations include keyword matching, ranking of search results, and optimization of results based on sentiment data.

[1799] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1800] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1801] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1802] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1803] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1804] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1805] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1806] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1807] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1808] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1809] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1810] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1811] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1812] 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.

[1813] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1814] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1815] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1816] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1817] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1818] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1819] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[1820] The following is further disclosed regarding the embodiments described above.

[1821] (Claim 1)

[1822] Means of receiving documents,

[1823] A means of generating a summary from a document,

[1824] A method for extracting important keywords from a summary,

[1825] A method for tagging documents based on extracted keywords,

[1826] Means of classifying documents,

[1827] Means of displaying tags,

[1828] A means of searching for documents classified based on the searched keywords,

[1829] A system that includes this.

[1830] (Claim 2)

[1831] The system according to claim 1, wherein the means for generating a summary from a document utilizes natural language processing technology.

[1832] (Claim 3)

[1833] The system according to claim 1, wherein the means for tagging documents is to automatically generate tags based on extracted keywords.

[1834] "Example 1"

[1835] (Claim 1)

[1836] Means of receiving documents,

[1837] A means of generating a summary from a document,

[1838] A method for extracting important keywords from a summary,

[1839] A method for tagging documents based on extracted keywords,

[1840] A means of displaying tags on the device,

[1841] A means of classifying documents based on tags,

[1842] A means of searching for documents classified based on the searched keywords,

[1843] A system that includes this.

[1844] (Claim 2)

[1845] The system according to claim 1, wherein the means for generating a summary from a document utilizes natural language processing technology.

[1846] (Claim 3)

[1847] The system according to claim 1, wherein the means for tagging documents is to automatically generate tags based on extracted keywords.

[1848] "Application Example 1"

[1849] (Claim 1)

[1850] Means of receiving documents,

[1851] A means of generating a summary from a document,

[1852] A method for extracting important keywords from a summary,

[1853] A method for tagging documents based on extracted keywords,

[1854] Means of classifying documents,

[1855] Means of displaying tags,

[1856] A means of searching for documents classified based on the searched keywords,

[1857] A means of reading documents via a smart device and uploading them to the cloud,

[1858] A means of displaying search results on the display of a smart device,

[1859] A system that includes this.

[1860] (Claim 2)

[1861] The system according to claim 1, wherein the means for generating a summary from a document utilizes natural language processing technology.

[1862] (Claim 3)

[1863] The system according to claim 1, wherein the means for tagging documents is to automatically generate tags based on extracted keywords.

[1864] "Example 2 of combining an emotion engine"

[1865] (Claim 1)

[1866] Means of receiving documents,

[1867] A means of generating a summary from a document,

[1868] A method for extracting important keywords from a summary,

[1869] A method for tagging documents based on extracted keywords,

[1870] Means for acquiring user sentiment data,

[1871] A means of adjusting the classification of tags and documents based on acquired sentiment data,

[1872] Means of classifying documents,

[1873] Means of displaying tags,

[1874] A means of searching for documents classified based on the searched keywords,

[1875] A system that includes this.

[1876] (Claim 2)

[1877] The system according to claim 1, wherein the means for generating a summary from a document utilizes natural language processing technology.

[1878] (Claim 3)

[1879] The system according to claim 1, wherein the means for tagging documents is to automatically generate tags based on extracted keywords.

[1880] "Application example 2 when combining with an emotional engine"

[1881] (Claim 1)

[1882] Means of receiving documents,

[1883] A means of generating a summary from a document,

[1884] A method for extracting important keywords from a summary,

[1885] A method for tagging documents based on extracted keywords,

[1886] A method of classifying documents using tags,

[1887] A means of generating user sentiment data,

[1888] A means of adjusting document tags based on sentiment data,

[1889] A means of displaying the generated tags,

[1890] A means for searching for documents categorized based on searched keywords and adjusted tags,

[1891] A system that includes this.

[1892] (Claim 2)

[1893] The system according to claim 1, wherein the means for generating a summary from a document utilizes natural language processing technology.

[1894] (Claim 3)

[1895] The system according to claim 1, wherein the means for tagging documents is to automatically generate tags based on extracted keywords. [Explanation of Symbols]

[1896] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. Means of receiving documents, A means of generating a summary from a document, A method for extracting important keywords from a summary, A method for tagging documents based on extracted keywords, Means of classifying documents, Means of displaying tags, A means of searching for documents classified based on the searched keywords, A system that includes this.

2. The system according to claim 1, wherein the means for generating a summary from a document utilizes natural language processing technology.

3. The system according to claim 1, wherein the means for tagging documents is to automatically generate tags based on extracted keywords.

Images