system

Abstract

Provide a system. An input means for inputting inquiry content; a refinement means for refining the query content received from the input means using a generative model; a display means for displaying the inquiry content revised by the revision means; A system including:

Description

[Technical Field]

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

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

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

[0004] A problem with inquiry work is that much of the time is spent exchanging unclear questions. Unclear questions take time to resolve, leading to inefficiency in work. The challenge is to solve this problem and clarify the content of inquiries and improve the efficiency of the overall work process. In addition, since many inquiries contain different, unique terminology and knowledge, technology that can address this is required. [Means for solving the problem]

[0005] The system of the present invention provides a mechanism for inputting a query and automatically revising it using a generative model. The system includes an input means, a revision means, a display means, and a generative model including embedding technology. A user inputs the query using the input means, and the content revised by the revision means is displayed on the display means. The system also includes a function that allows the user to interactively select or modify revision suggestions. This can improve the clarity of the query and enhance work efficiency.

[0006] "Inquiry content" is a description of the problem or question the user wants to solve.

[0007] The "input means" is an interface through which the user inputs the contents of an inquiry.

[0008] A "generative model" is a machine learning model that refines query content based on natural language processing.

[0009] The "refinement means" is a function that automatically corrects and improves the query content using a generative model.

[0010] The "display means" is an interface for displaying the refined inquiry content to the user.

[0011] "Embedding technology" is a technology used in generative models to accommodate specific knowledge and terminology.

[0012] "Interactive means" are features that allow the user to select or modify revision suggestions. [Brief explanation of the drawings]

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

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

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

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

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

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

[0019] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0020] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0021] [First embodiment]

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

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

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

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

[0026] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

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

[0028] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

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

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

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

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

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

[0034] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically revise query content entered by a user using a generative model and convert it into a clear query content.

[0035] System configuration

[0036] Server side

[0037] The server receives the query and generates a refinement proposal using a generative model. The generative model generates text to clarify the question based on natural language processing. For example, if a user enters "Please tell me the shipping status of my item," the server inputs this question into the generative model and generates a refinement proposal such as "Please tell me in detail the shipping status of the item I currently ordered."

[0038] The server hosts a pre-trained model, for example using the Transformers library, as a generative model. The server tokenizes the acquired question, inputs it into the generative model, and outputs the revised sentence. The server then returns this revised sentence in JSON format to the user's device.

[0039] Terminal side

[0040] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript (registered trademark) on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[0041] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision suggestion" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me in detail about the shipping status of the product I have currently ordered," the terminal will display this revision suggestion on the web page.

[0042] User side

[0043] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0044] Specific examples

[0045] Specific examples are shown below.

[0046] The user enters "Please tell me the shipping status of my item" into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the item I have currently ordered," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and modify the original question.

[0047] As a result, the system of the present invention can improve the clarity of the inquiry content and improve the efficiency of the business process.

[0048] The processing flow will be explained below.

[0049] Step 1:

[0050] The user inputs a question into the inquiry form.

[0051] The user enters "Please tell me the shipping status of my item" into the text area of ​​the web form.

[0052] Step 2:

[0053] The user clicks the "Get Revision Suggestions" button.

[0054] The user clicks a button to request revision of the inquiry.

[0055] Step 3:

[0056] The terminal obtains the user's input.

[0057] The terminal uses JavaScript to obtain the question "Please tell me the shipping status of the product" entered in the text area.

[0058] Code such as const questionText = document.getElementById("question").value; is executed.

[0059] Step 4:

[0060] The terminal sends a request to the server.

[0061] The device uses the fetch API to send a POST request containing the question to the server.

[0062] Code such as const response = await fetch('http: / / server_address / suggestions', {...}); is executed.

[0063] Step 5:

[0064] The server receives the request.

[0065] The server receives a POST request from the user terminal. The question is included in the request body.

[0066] Step 6:

[0067] The server uses the generative model to refine the text.

[0068] The server uses an AI model to tokenize the question, input it into the model, and generate a refined sentence.

[0069] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[0070] Step 7:

[0071] The server decodes the generated revisions and converts them into text.

[0072] The server decodes the generated token and converts it into human-readable text.

[0073] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[0074] Step 8:

[0075] The server returns a response including a revision proposal to the terminal.

[0076] The server includes the generated revision proposal in a response in JSON format and sends it back to the terminal.

[0077] Step 9:

[0078] The terminal receives the response from the server.

[0079] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[0080] Code such as const data = await response.json(); is executed.

[0081] Step 10:

[0082] The terminal displays the revised draft on the screen.

[0083] The terminal displays the received revision proposal in a specified HTML element on a web page.

[0084] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[0085] Step 11:

[0086] The user checks the revised draft.

[0087] The user views the revised question displayed on the screen, "Please tell me in detail about the shipping status of the item I have currently ordered," and compares it with the original question.

[0088] Step 12:

[0089] The user can modify the question as needed.

[0090] The user uses the revised suggestions as a reference and modifies the original question to something like, "Please tell me in detail about the shipping status of the item I have currently ordered."

[0091] These steps allow users to formulate clear and specific inquiries, improving work efficiency.

[0092] Example 1

[0093] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0094] In today's information society, users make numerous inquiries, but these are often expressed vaguely or unclearly. This makes it difficult for the recipient (e.g., a company or support staff) to accurately understand the intent, and it can take time to respond. Furthermore, if the inquiry content is unclear, the quality of the response will decline, which in turn reduces user satisfaction. To solve this problem, a system is needed that can automatically refine the inquiry content and convert it into clear and specific content.

[0095] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0096] In this invention, the server includes an input means for inputting an inquiry, a refinement means for refining the inquiry received from the input means using a generative model, a processing means for tokenizing the inquiry for input to the generative model, a return means for returning the inquiry refined by the refinement means in JSON format, and a display means for displaying the inquiry refined by the refinement means. This automatically converts the inquiry input by the user into clear and specific content, making it possible to improve the efficiency of inquiry operations and the quality of responses.

[0097] "Inquiry content" refers to a question or request for information that a user sends to the system through an input means.

[0098] "Input means" refers to a device or software that provides an interface for a user to input inquiry details.

[0099] A "generative model" refers to an algorithm or machine learning model that automatically generates refinement suggestions based on the specified inquiry content.

[0100] "Refining means" refers to the process or function that uses a generative model to refine the content of a received query for clarity.

[0101] "Processing means" refers to the process or function that converts the query content into an appropriate form before inputting it into the generative model.

[0102] "JSON format" stands for JavaScript Object Notation and refers to a lightweight data exchange format for structuring data.

[0103] The "returning means" refers to a process or function for returning the generated revision proposal to the user or terminal.

[0104] "Display means" refers to a device or software for displaying the refined inquiry content on a terminal in a format that can be confirmed by the user.

[0105] "Embedding technology" refers to a data representation technique that enables generative models to correspond to specific knowledge and terminology.

[0106] "Interactive means" refers to a device or software that provides an interface to allow a user to select or modify a refinement.

[0107] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically refine query content entered by a user using a generative model and convert it into an unambiguous query content.

[0108] Server side

[0109] The server receives the query and generates a refinement proposal using a generative model. This generative model uses natural language processing. Specifically, the server processes the query using a tokenization processing means and generates a refinement proposal.

[0110] The server hosts a pre-trained generative model using the Transformers library, a powerful tool implemented in Python and specialized for natural language processing. The server takes the user's question, tokenizes it, and feeds it into the generative model. The generative model generates clearer text and returns its refinements in JSON format.

[0111] For example, if a user inputs an inquiry such as "Please tell me the shipping status of the product," the server inputs this information into the generative model and generates a refinement proposal such as "Please tell me in detail about the shipping status of the product I have currently ordered."

[0112] Terminal side

[0113] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying revision suggestions. Specifically, the terminal has a web form where the user enters their inquiry. When the user clicks the "Get revision suggestions" button to submit the input, a request is sent to the server using JavaScript.

[0114] The terminal receives the response from the server and displays the received revision suggestions on the screen, allowing the user to check the revision suggestions and revise the original question.

[0115] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision proposal" button, the terminal will send this inquiry to the server. If the server returns a revision proposal saying "Please tell me the details of the shipping status of the product I have currently ordered," the terminal will display this revision proposal on the web page.

[0116] User side

[0117] The user begins operating the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the server displays the revision suggestions on the device, the user can use these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0118] Specific examples

[0119] A concrete example is given below. A user enters "Please tell me the shipping status of my product" into an inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the product I have currently ordered," and sends it back to the device. The device displays this revision suggestion on a web page. The user checks the displayed revision suggestion and modifies the original question.

[0120] Prompt Sentence Examples

[0121] Below are some examples of prompts to input to a generative AI model.

[0122] "Please make the sentence 'Please let me know the shipping status of the item' clearer."

[0123] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0124] Step 1:

[0125] The user enters the inquiry into a web form on the device. The input is text data, specifically a question such as "Please tell me the shipping status of my item." The user finishes entering the content into the web form as the output of this step.

[0126] Step 2:

[0127] The user clicks the "Get revision proposal" button on the device. This action causes the device to send the query to the server using JavaScript. The input data is the question entered by the user, and the output data is the request data in JSON format sent to the server.

[0128] Step 3:

[0129] The server receives the query sent from the terminal. It receives the question sent in JSON format as input. Next, it converts the question into tokens using a tokenization processing means. This converts the question into a format that can be input to the generative model. The server outputs the query in token format.

[0130] Step 4:

[0131] The server invokes the generative model and provides the tokenized question as input. The generative model generates clear refinement suggestions based on this input. For example, in response to the input "Please tell me the shipping status of my item," the model outputs the refinement suggestion "Please tell me in detail about the shipping status of the item I have currently ordered."

[0132] Step 5:

[0133] The server converts the generated revisions into JSON format and sends them back to the terminal. At this time, the input is the revisions output by the generative model, and the output is JSON format data.

[0134] Step 6:

[0135] The terminal receives the JSON data returned from the server. The input is the JSON data containing the revision suggestions returned from the server. The terminal parses this data and converts it into a format suitable for display on a web page. For example, JavaScript can be used to display the revision suggestions in HTML elements.

[0136] Step 7:

[0137] The terminal displays the converted revised version to the user. The user checks the displayed version and revises the question if necessary. The input is the converted revised version, and the output at this step is a clear query for the user to confirm.

[0138] Through this series of processing steps, the system of the present invention can automatically refine the content of the user's inquiry and convert it into clearer and more specific content.

[0139] (Application example 1)

[0140] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0141] When users make inquiries to customer support on online shopping sites, the content of their inquiries is often unclear, resulting in delayed responses. Furthermore, users have difficulty formulating appropriate inquiries, which reduces the efficiency of inquiry processes. To solve this problem, a method for making inquiries clear and specific is needed.

[0142] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0143] In this invention, the server includes an input means for inputting a query, a revision means for revising the query received from the input means using a generative model, a display means for displaying the query revised by the revision means, and a transmission means for confirming the displayed revision plan and transmitting the query. This enables a user to easily create a clear and specific query and receive a prompt and appropriate response.

[0144] The "inquiry content" is a specific description of the question or request that the user wants customer support to resolve.

[0145] The "input means" is a device or software that provides an interface for a user to input inquiry content to the system.

[0146] A "generative model" is an algorithm or program based on natural language processing that converts input queries into clearer, more specific sentences.

[0147] The "refining means" is a device or software that automatically refines the inquiry content received from the input means using a generative model and converts it into more appropriate inquiry content.

[0148] The "display means" is a device or software for visually presenting the inquiry content revised by the revision means to the user.

[0149] "Transmission means" refers to a device or software for transmitting the user-confirmed revision proposal to customer support.

[0150] This invention relates to a system for improving the efficiency of user inquiries on online shopping sites. The system automatically refines the content of inquiries entered by users and converts them into clear and specific inquiries. A specific embodiment of this system is described in detail below.

[0151] System configuration

[0152] Server side

[0153] The server receives the query and generates refinement proposals using the generative model. Specifically, it uses the following hardware and software:

[0154] Hardware:

[0155] General server computer

[0156] software:

[0157] Natural Language Processing Generative Model: Transformers (Hugging Face)

[0158] Programming language: Python

[0159] Web server: Node.js

[0160] Database: MongoDB

[0161] The server first tokenizes the query received from the user and inputs it into a generative model. The generative model then uses natural language processing technology to generate clearer and more specific query sentences. For example, if a user inputs "Please tell me the shipping status of my item," the server inputs this sentence into the generative model and generates a refinement suggestion: "Please tell me in detail the shipping status of the item I currently ordered." The server then returns this refinement suggestion to the terminal in JSON format.

[0162] Terminal side

[0163] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. Specifically, the following hardware and software are used:

[0164] Hardware:

[0165] Smartphone

[0166] software:

[0167] Frontend: React Native

[0168] API communication: Axios

[0169] The device provides a user interface (UI) for entering inquiries using a web form. When the user clicks the "Get revision suggestions" button, a request is sent to the server using JavaScript in the browser. After the server returns the revision suggestions, the device displays them on the screen.

[0170] User side

[0171] The user operates the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the device, the user can refer to the suggestions to revise the inquiry and click the send button to send it to customer support.

[0172] Specific examples

[0173] For example, if a user enters "Please tell me the shipping status of the product" into the inquiry form and clicks the "Get revision proposal" button, the following process will occur.

[0174] 1. User input

[0175] Input: "Please let me know the shipping status of the item."

[0176] 2. The server generates a revision plan

[0177] Suggested revision: "Please tell me the details of the shipping status of the item I ordered."

[0178] 3. The device will display the revised version.

[0179] Message: "Please tell me the details of the shipping status of the item I ordered."

[0180] 4. User reviews the revision and submits an inquiry

[0181] In this way, users can easily formulate clear and specific questions and receive prompt and appropriate responses.

[0182] This system will improve the efficiency of customer support operations and also improve the user experience.

[0183] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0184] Step 1:

[0185] The user enters a question into the inquiry form. The terminal acquires the input and temporarily stores it. The input in this case is "Please tell me the shipping status of the product."

[0186] Step 2:

[0187] The device sends the entered query to the server, packaging it in JSON format and sending it to the server's API endpoint using an HTTP POST request.

[0188] Step 3:

[0189] The server retrieves the received JSON data and extracts the query content. Here, it parses the data and converts it into a format that can be processed by the generative model.

[0190] Step 4:

[0191] The server inputs the query into the generative model and generates refinement suggestions. The generative model performs tokenization and appropriate embeddings to generate clear sentences. For example, the refinement suggestion generated is, "Please tell me the details of the shipping status of the item I've currently ordered."

[0192] Step 5:

[0193] The server packages the generated revision suggestions in JSON format and returns them to the terminal, sending JSON data containing the revision suggestions as an HTTP response.

[0194] Step 6:

[0195] The device receives the revision proposals from the server and displays them on the user interface by parsing the received JSON data, extracting the revision proposals, and displaying them on the screen.

[0196] Step 7:

[0197] The user checks the revised proposal displayed on the terminal, makes corrections as necessary, then confirms the revised inquiry content and clicks the send button to send it to customer support.

[0198] Step 8:

[0199] The device sends the inquiry confirmed by the user to the server again as the final inquiry, again in JSON format.

[0200] Step 9:

[0201] The server receives the final query and forwards it to the customer support team, where it is stored in a database and a response is initiated.

[0202] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0203] This invention is a system that combines a series of processes for inputting, revising, and displaying query content with an emotion engine that recognizes and analyzes user emotions. The system aims to generate more appropriate query sentences by automatically revising the query content entered by the user using a generative model and further analyzing the user's emotions.

[0204] System configuration

[0205] Server side

[0206] The server receives the query and generates refinement suggestions using a generative model. It also has the function of analyzing the emotions contained in the user's query using an emotion engine. The generative model generates text to clarify the question based on natural language processing.

[0207] For example, if a user enters "My product delivery is late, please tell me what's going on," the server inputs this question into the generative model and generates a refinement suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why the delivery is late." The emotion engine also analyzes the user's emotions from this question and recognizes emotions such as "dissatisfaction" and "irritation." Based on this information, the refinement suggestion is adjusted to generate a query that is more suited to the user.

[0208] The server hosts pre-trained models, e.g., using the Transformers library, as generative models, and also uses pre-trained emotion recognition models as emotion engines.

[0209] Terminal side

[0210] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[0211] For example, if a user inputs "The delivery of my product is late, please let me know what's going on" and clicks the "Get revision suggestions" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late," the terminal will display this revision suggestion on the web page.

[0212] User side

[0213] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0214] Specific examples

[0215] Specific examples are shown below.

[0216] The user enters "The delivery of my product is late, so please let me know what's going on." into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed." The emotion engine also analyzes emotions such as "dissatisfaction" and "irritation" from the question. Based on this information, the server adjusts the revision suggestion and generates a revision suggestion that takes the user's emotions into consideration, such as "I'm sorry to have kept you waiting. Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and revise the original question.

[0217] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[0218] The processing flow will be explained below.

[0219] Step 1:

[0220] The user inputs a question into the inquiry form.

[0221] The user types "My item is taking a long time to ship, please let me know what's going on" into a text area on a web form.

[0222] Step 2:

[0223] The user clicks the "Get Revision Suggestions" button.

[0224] The user clicks a button to request revision of the inquiry.

[0225] Step 3:

[0226] The terminal obtains the user's input.

[0227] The terminal uses JavaScript to obtain the question entered in the text area, "The shipment of the product is late, please let me know what's going on."

[0228] Code such as const questionText = document.getElementById("question").value; is executed.

[0229] Step 4:

[0230] The terminal uses the emotion engine to send a request to the server to recognize the user's emotion.

[0231] The device creates a request to send the question to the emotion engine for emotion analysis.

[0232] Code such as const emotionResponse = await fetch('http: / / server_address / emotion', {...}); is executed.

[0233] Step 5:

[0234] The server receives the query.

[0235] The server receives the POST request from the user terminal and analyzes the question.

[0236] Step 6:

[0237] The server uses an emotion engine to analyze the user's emotions.

[0238] The server inputs the question into an emotion engine and obtains emotional information such as "dissatisfaction" and "irritation."

[0239] Step 7:

[0240] The server uses the generative model to refine the text.

[0241] The server uses an AI generation model to tokenize the question, input it into the model, and generate a refined sentence.

[0242] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[0243] Step 8:

[0244] The server decodes the generated revisions and converts them into text.

[0245] The server decodes the generated token and converts it into human-readable text.

[0246] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[0247] Step 9:

[0248] The server takes into account the user's emotional information and adjusts the revision proposal.

[0249] Based on the results of the emotion engine, the server adds emotional considerations to the revised proposal, for example adding phrases such as "Sorry for keeping you waiting."

[0250] Step 10:

[0251] The server returns a response including a revision proposal to the terminal.

[0252] The server includes the adjusted revision proposal in a JSON format response and sends it back to the terminal.

[0253] Step 11:

[0254] The terminal receives the response from the server.

[0255] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[0256] Code such as const data = await response.json(); is executed.

[0257] Step 12:

[0258] The terminal displays the revised draft on the screen.

[0259] The terminal displays the received revision proposal in a specified HTML element on a web page.

[0260] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[0261] Step 13:

[0262] The user checks the revised draft.

[0263] The user views the revised question displayed on the screen, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know the reason for the delay in shipping." and compares it with the original question.

[0264] Step 14:

[0265] The user can modify the question as needed.

[0266] The user uses the revised suggestions as a reference and modifies the original question to something like, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[0267] These steps allow users to create clear, specific inquiries that are sensitive to emotions, improving work efficiency.

[0268] Example 2

[0269] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0270] In conventional inquiry systems, the inquiries entered by users are not always clear, or do not reflect the user's feelings, resulting in inappropriate inquiries. This results in poor efficiency in responding to inquiries and low user satisfaction. In particular, mechanical revisions that ignore the user's feelings can lead to poor communication quality and even cause problems. Therefore, there is a need for a system that can make the inquiries entered by users clearer and more appropriate, and that can respond in a way that takes the user's feelings into consideration.

[0271] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0272] In this invention, the server includes an input means for inputting the content of a query, a refining means for refining the received content of the query using a generative model, a display means for displaying the refined content of the query, an emotion recognition means for analyzing emotions from the generated content of the query, and an adjustment means for adjusting the content of the query based on the analyzed emotions. This makes it possible to make the content of the user's query clear and appropriate, and to generate an optimal query sentence that takes the user's emotions into consideration.

[0273] A "query" is the text a user enters to request specific information.

[0274] "Input means" refers to an interface or device that a user uses to input the contents of a query.

[0275] A "generative model" is an algorithm or software that uses natural language processing techniques to automatically refine input text.

[0276] The "elaboration means" refers to a means for modifying the input query content to make it clearer and more appropriate using a generative model.

[0277] "Display means" refers to an interface or device for visually presenting the refined inquiry content to the user.

[0278] "Emotion recognition means" refers to an algorithm or software that analyzes and recognizes a user's emotions from the content of the input inquiry.

[0279] The "adjustment means" refers to a means for optimizing the inquiry content based on the emotion information analyzed by the emotion recognition means.

[0280] "Interactive means" refers to an input interface or device that allows the user to review and modify the final query statement generated.

[0281] "Embedding techniques" refer to techniques used to make generative models correspond to specific knowledge and vocabulary.

[0282] The present invention is a system that combines a series of processes for inputting, revising, and displaying the content of a query with an emotion engine that recognizes and analyzes the user's emotions. Specific embodiments for carrying out the present invention will be described below.

[0283] Server side

[0284] The server receives the query content, generates refinement proposals using the generative model, and analyzes the emotions contained in the user's query content using an emotion engine.

[0285] The software used is the Transformers library, which is based on natural language processing technology, for the generative model. A pre-trained generative AI model is hosted on a server and inputs the query content sent by the user.

[0286] As a concrete example, if a user inputs "The delivery of my product is late, please let me know what's going on," the server will run this question through the generative model and generate a refinement suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late."

[0287] The server then uses an emotion recognition engine to recognize the emotions contained in the user's query. The emotion engine uses a pre-trained emotion recognition model that analyzes emotions from text and identifies emotions such as "frustration" or "irritation."

[0288] Based on this information, the server generates a final query, such as "I'm sorry for the wait. Please let me know the status of the delivery of my order. I'd also like to know why the delivery is delayed."

[0289] Terminal side

[0290] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. A web form is used as the specific user interface.

[0291] This involves the user entering their query in a text field and clicking the "Get Revision Suggestions" button. The device then uses JavaScript in the browser to send a request to the server. Specifically, it uses an Ajax request to send data asynchronously.

[0292] Once the response from the server is received, the device displays the revised version on the web page. For example, if the server returns a message like, "Sorry for the wait. Please let me know the status of the delivery of the item I ordered. I would also like to know why the delivery is delayed," this message will be displayed on the screen.

[0293] User side

[0294] The user enters a question into the inquiry form on the device and clicks the "Get revision proposal" button. This operation sends the inquiry to the server.

[0295] When the server returns the revised query, the content is displayed on the terminal. The user can review it and make corrections as necessary, which allows the user to create a clearer and more appropriate query.

[0296] Specific examples

[0297] Specific examples are shown below.

[0298] 1. The user enters "The product delivery is late, please let me know what's going on" into the inquiry form and clicks the "Get revision suggestions" button.

[0299] 2. The device sends this input to the server.

[0300] 3. The server uses the generative model to generate a refinement suggestion such as, "Please tell me the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[0301] 4. The emotion engine analyzes the input for emotions such as "dissatisfaction" or "irritation." Based on this information, it adjusts the suggested revision and generates a message like, "Sorry for the wait. Please let me know the status of the delivery of my order. I would also like to know why the delivery is delayed."

[0302] 5. The final query is sent back to the terminal, which displays it on the screen.

[0303] 6. The user checks the displayed revision suggestions and corrects the original question.

[0304] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[0305] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0306] Step 1: User enters inquiry

[0307] The user accesses a web form on the device and enters the inquiry in the text field. Specifically, the user opens the form in a browser and enters the prompt sentence, "The delivery of my product is late. Please let me know what's going on."

[0308] Input: Inquiry details

[0309] Output: The inquiry content is saved on the terminal.

[0310] Step 2: The device sends the input to the server

[0311] When the user clicks the "Get Revision Suggestions" button, the device uses JavaScript to send the query to the server. This operation is performed asynchronously using an Ajax request with the HTTP POST method.

[0312] Input: Inquiry details

[0313] Output: The query is sent to the server

[0314] Step 3: The server inputs the query into the generative model

[0315] The server passes the query received from the device through a generative model, which uses the Transformers library, a natural language processing technology, to generate clear refinement suggestions based on the query.

[0316] For example, the server converts an inquiry received such as "The shipment of the product is late, please tell me what's going on" into a refined proposal such as "Please tell me the shipping status of the product I ordered. I would also like to know the reason for the delay in shipping."

[0317] Input: Inquiry details

[0318] Data processing: Using generative AI models to refine queries

[0319] Output: Elaboration plan

[0320] Step 4: The server analyzes the emotion using the emotion recognition method.

[0321] The server inputs the generated revision plan into an emotion recognition means for analyzing the user's emotion, which uses an emotion recognition model to extract emotion information from the revision plan.

[0322] For example, emotions such as "dissatisfaction" and "irritation" are recognized as analysis results.

[0323] Input: Revision

[0324] Data Computing: Analyzing emotions using emotion recognition models

[0325] Output: Emotional information

[0326] Step 5: The server generates the final query based on the emotion information.

[0327] The server combines the generated refinement proposals with the emotion information to generate a final query that takes the user's emotions into consideration.

[0328] For example, you might say, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[0329] Input: Revision plan, emotional information

[0330] Data processing: Adjusting query sentences based on emotional information

[0331] Output: Final query

[0332] Step 6: The server sends the final query back to the terminal.

[0333] After the final query has been generated, the server sends it back to the terminal, which receives it and prepares it for display to the user.

[0334] Input: Final query

[0335] Output: The final query is sent to the terminal

[0336] Step 7: The terminal displays the final query

[0337] The final query returned by the server is displayed on a web page on the device, which uses HTML and JavaScript to display the text in a user-friendly way.

[0338] For example, the browser screen might display, "We apologize for the wait. Please let us know the shipping status of the item you ordered. We would also like to know the reason for the delay in shipping."

[0339] Input: Final query

[0340] Output: The query is displayed on the screen.

[0341] Step 8: User confirms and modifies the query

[0342] The user can review the final query displayed and make corrections as necessary, allowing the user to create a clear and appropriate query and proceed to final submission.

[0343] For example, the user confirms or corrects by saying, "Please let me know the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[0344] Input: Final query

[0345] Output: Modified query

[0346] (Application example 2)

[0347] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0348] When responding to inquiries in physical stores, it can be difficult for customers to enter appropriate and clear inquiry content. It is also necessary to properly recognize customer emotions and respond accordingly, but current systems are not sufficient in this regard. In particular, responding without understanding customer emotions carries the risk of lowering the quality of service. Facing these challenges, there is a need for the development of a system that allows customers to make inquiries in stores effectively and with consideration for their emotions.

[0349] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0350] In this invention, the server includes an input means for inputting an inquiry, a refining means for refining the inquiry received from the input means using a generative model, a display means for displaying the inquiry refined by the refining means, an emotion recognition means for analyzing the emotion of the inquiry, and an optimization means for optimizing the inquiry based on the emotion analyzed by the emotion recognition means. This makes it possible to clarify the inquiry input by a customer and to provide an optimized inquiry sentence that takes emotion into consideration.

[0351] "Input means" refers to a device or interface that allows a user to input inquiry details.

[0352] "Refining means" refers to a function that uses a generative model to revise the query entered by the user into a clearer and more specific sentence.

[0353] "Display means" refers to a device or interface for visually presenting the refined query content or optimized query sentence to the user.

[0354] "Emotion recognition means" refers to a function that analyzes emotions from the inquiry content entered by the user and identifies those emotions.

[0355] The "optimization means" refers to a function that appropriately adjusts the content of the inquiry based on the emotions analyzed by the emotion recognition means, and generates an inquiry statement that takes the user's emotions into consideration.

[0356] "Generative model" refers to a machine learning model that uses natural language processing techniques to generate and modify text.

[0357] This invention is a system that automatically refines customer inquiries and analyzes user sentiment to improve customer support in brick-and-mortar stores. The system consists of three main components: a server, a terminal, and a user.

[0358] Server side

[0359] The server receives the query content and processes it using a generative model and emotion recognition means.

[0360] 1. Generative Model

[0361] Hardware used: High-performance server computer

[0362] Software used: Natural language processing model using the Transformers library

[0363] Data computation: The received query content is input into the generative model to generate clear refinement suggestions.

[0364] 2. Emotion recognition means

[0365] Hardware used: High-performance server computer

[0366] Software used: Sentiment analysis model

[0367] Data calculation: Analyze emotions from inquiries and identify feelings such as "dissatisfaction" or "irritation."

[0368] 3. Optimization Methods

[0369] An optimized query is generated based on the refinement suggestions and the sentiment analysis results.

[0370] As a concrete example, if a user inputs "The delivery of my item is late, please let me know what's going on," the server inputs this into the generative model. The generative model generates a refinement suggestion such as "Please tell me the delivery status of the item I ordered. I would also like to know why the delivery is delayed." At the same time, the sentiment analysis model identifies emotions such as "dissatisfaction" and "irritation." Through optimization, the final sentence generated is "I'm sorry to have kept you waiting. Please let me know the delivery status of the item I ordered. I would also like to know why the delivery is delayed."

[0371] Terminal side

[0372] The terminal has the role of transmitting the inquiry content entered by the user to the server and displaying the response from the server.

[0373] 1. Interface

[0374] Hardware used: Smartphone, tablet

[0375] Software used: Web application (HTML, JavaScript)

[0376] Data Entry: User enters their query into the text field and presses the "Get Revisions" button

[0377] 2. Data transmission and reception

[0378] Software used: Data communication using AJAX

[0379] Data calculation: Sends the input query to the server and receives refinement suggestions and sentiment analysis results.

[0380] As a concrete example, a user inputs "The delivery of my product is late, please let me know what's going on," and clicks the "Get revision proposal" button. The device sends this content to the server, which returns a revision proposal saying, "I'm sorry to have kept you waiting. Please let me know the delivery status of the product I ordered. I'd also like to know why the delivery is late."

[0381] User side

[0382] The user is responsible for inputting and modifying the query using a terminal and receiving the optimized query statement from the server.

[0383] 1. Input

[0384] The user enters their inquiry into the text field.

[0385] 2. Correction and confirmation

[0386] Check the optimized query and modify it if necessary

[0387] For example, a user can input "The shipment of my item is delayed, so please let me know what's going on." and then review the generated sentence "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know why the shipment is delayed." and correct the original inquiry.

[0388] The present invention improves the efficiency of responding to inquiries in physical stores and also realizes optimal responses that take into consideration the feelings of users.

[0389] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0390] Step 1:

[0391] The user inputs the inquiry into the terminal interface. Specifically, the user inputs the inquiry into the text field and clicks the "Get revision proposal" button. The input data is the inquiry in text format.

[0392] Step 2:

[0393] The terminal sends the entered query content to the server. In this process, the input content is sent to the server using asynchronous communication using AJAX. The input data is the text content entered by the user and is sent to the server as output.

[0394] Step 3:

[0395] The server inputs the query received into a generative model to generate a refinement proposal. Specifically, it uses the Transformers library to refine the query into a clearer and more specific sentence. The input data is the user's query, and the output data is the text refined by the generative model.

[0396] Step 4:

[0397] The server inputs the refined query content into the emotion recognition means to analyze the emotion. An emotion analysis model is used to extract emotions such as "dissatisfaction" or "irritation" from the text. The input data is the refined query sentence, and the output data is the analyzed emotion information.

[0398] Step 5:

[0399] The server optimizes the query content based on the emotion recognition results. Specifically, it adjusts the wording to be appropriate, taking into account the emotion information. The input data is the emotion analysis results and the refined query, and the output data is the optimized query.

[0400] Step 6:

[0401] The server sends the optimized query statement to the terminal. In this process, the generated optimized statement is returned to the terminal. The input data is the optimized query statement, and it is sent to the terminal as output data.

[0402] Step 7:

[0403] The terminal displays the optimized query received by the terminal to the user. Specifically, the query received from the server is displayed on the interface so that the user can confirm and modify it. The input data is the optimized query, and the output data is the text content displayed to the user.

[0404] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0405] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0406] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0407] [Second embodiment]

[0408] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0409] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0410] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0411] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0412] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0413] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0414] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0415] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0416] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0417] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0418] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0419] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0420] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically revise query content entered by a user using a generative model and convert it into a clear query content.

[0421] System configuration

[0422] Server side

[0423] The server receives the query and generates a refinement proposal using a generative model. The generative model generates text to clarify the question based on natural language processing. For example, if a user enters "Please tell me the shipping status of my item," the server inputs this question into the generative model and generates a refinement proposal such as "Please tell me in detail the shipping status of the item I currently ordered."

[0424] The server hosts a pre-trained model, for example using the Transformers library, as a generative model. The server tokenizes the acquired question, inputs it into the generative model, and outputs the revised sentence. The server then returns this revised sentence in JSON format to the user's device.

[0425] Terminal side

[0426] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[0427] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision suggestion" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me in detail about the shipping status of the product I have currently ordered," the terminal will display this revision suggestion on the web page.

[0428] User side

[0429] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0430] Specific examples

[0431] Specific examples are shown below.

[0432] The user enters "Please tell me the shipping status of my item" into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the item I have currently ordered," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and modify the original question.

[0433] As a result, the system of the present invention can improve the clarity of the inquiry content and improve the efficiency of the business process.

[0434] The processing flow will be explained below.

[0435] Step 1:

[0436] The user inputs a question into the inquiry form.

[0437] The user enters "Please tell me the shipping status of my item" into the text area of ​​the web form.

[0438] Step 2:

[0439] The user clicks the "Get Revision Suggestions" button.

[0440] The user clicks a button to request revision of the inquiry.

[0441] Step 3:

[0442] The terminal obtains the user's input.

[0443] The terminal uses JavaScript to obtain the question "Please tell me the shipping status of the product" entered in the text area.

[0444] Code such as const questionText = document.getElementById("question").value; is executed.

[0445] Step 4:

[0446] The terminal sends a request to the server.

[0447] The device uses the fetch API to send a POST request containing the question to the server.

[0448] Code such as const response = await fetch('http: / / server_address / suggestions', {...}); is executed.

[0449] Step 5:

[0450] The server receives the request.

[0451] The server receives a POST request from the user terminal. The question is included in the request body.

[0452] Step 6:

[0453] The server uses the generative model to refine the text.

[0454] The server uses an AI model to tokenize the question, input it into the model, and generate a refined sentence.

[0455] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[0456] Step 7:

[0457] The server decodes the generated revisions and converts them into text.

[0458] The server decodes the generated token and converts it into human-readable text.

[0459] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[0460] Step 8:

[0461] The server returns a response including a revision proposal to the terminal.

[0462] The server includes the generated revision proposal in a response in JSON format and sends it back to the terminal.

[0463] Step 9:

[0464] The terminal receives the response from the server.

[0465] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[0466] Code such as const data = await response.json(); is executed.

[0467] Step 10:

[0468] The terminal displays the revised draft on the screen.

[0469] The terminal displays the received revision proposal in a specified HTML element on a web page.

[0470] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[0471] Step 11:

[0472] The user checks the revised draft.

[0473] The user views the revised question displayed on the screen, "Please tell me in detail about the shipping status of the item I have currently ordered," and compares it with the original question.

[0474] Step 12:

[0475] The user can modify the question as needed.

[0476] The user uses the revised suggestions as a reference and modifies the original question to something like, "Please tell me in detail about the shipping status of the item I have currently ordered."

[0477] These steps allow users to formulate clear and specific inquiries, improving work efficiency.

[0478] Example 1

[0479] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0480] In today's information society, users make numerous inquiries, but these are often expressed vaguely or unclearly. This makes it difficult for the recipient (e.g., a company or support staff) to accurately understand the intent, and it can take time to respond. Furthermore, if the inquiry content is unclear, the quality of the response will decline, which in turn reduces user satisfaction. To solve this problem, a system is needed that can automatically refine the inquiry content and convert it into clear and specific content.

[0481] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0482] In this invention, the server includes an input means for inputting an inquiry, a refinement means for refining the inquiry received from the input means using a generative model, a processing means for tokenizing the inquiry for input to the generative model, a return means for returning the inquiry refined by the refinement means in JSON format, and a display means for displaying the inquiry refined by the refinement means. This automatically converts the inquiry input by the user into clear and specific content, making it possible to improve the efficiency of inquiry operations and the quality of responses.

[0483] "Inquiry content" refers to a question or request for information that a user sends to the system through an input means.

[0484] "Input means" refers to a device or software that provides an interface for a user to input inquiry details.

[0485] A "generative model" refers to an algorithm or machine learning model that automatically generates refinement suggestions based on the specified inquiry content.

[0486] "Refining means" refers to the process or function that uses a generative model to refine the content of a received query for clarity.

[0487] "Processing means" refers to the process or function that converts the query content into an appropriate form before inputting it into the generative model.

[0488] "JSON format" stands for JavaScript Object Notation and refers to a lightweight data exchange format for structuring data.

[0489] The "returning means" refers to a process or function for returning the generated revision proposal to the user or terminal.

[0490] "Display means" refers to a device or software for displaying the refined inquiry content on a terminal in a format that can be confirmed by the user.

[0491] "Embedding technology" refers to a data representation technique that enables generative models to correspond to specific knowledge and terminology.

[0492] "Interactive means" refers to a device or software that provides an interface to allow a user to select or modify a refinement.

[0493] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically refine query content entered by a user using a generative model and convert it into an unambiguous query content.

[0494] Server side

[0495] The server receives the query and generates a refinement proposal using a generative model. This generative model uses natural language processing. Specifically, the server processes the query using a tokenization processing means and generates a refinement proposal.

[0496] The server hosts a pre-trained generative model using the Transformers library, a powerful tool implemented in Python and specialized for natural language processing. The server takes the user's question, tokenizes it, and feeds it into the generative model. The generative model generates clearer text and returns its refinements in JSON format.

[0497] For example, if a user inputs an inquiry such as "Please tell me the shipping status of the product," the server inputs this information into the generative model and generates a refinement proposal such as "Please tell me in detail about the shipping status of the product I have currently ordered."

[0498] Terminal side

[0499] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying revision suggestions. Specifically, the terminal has a web form where the user enters their inquiry. When the user clicks the "Get revision suggestions" button to submit the input, a request is sent to the server using JavaScript.

[0500] The terminal receives the response from the server and displays the received revision suggestions on the screen, allowing the user to check the revision suggestions and revise the original question.

[0501] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision proposal" button, the terminal will send this inquiry to the server. If the server returns a revision proposal saying "Please tell me the details of the shipping status of the product I have currently ordered," the terminal will display this revision proposal on the web page.

[0502] User side

[0503] The user begins operating the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the server displays the revision suggestions on the device, the user can use these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0504] Specific examples

[0505] A concrete example is given below. A user enters "Please tell me the shipping status of my product" into an inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the product I have currently ordered," and sends it back to the device. The device displays this revision suggestion on a web page. The user checks the displayed revision suggestion and modifies the original question.

[0506] Prompt Sentence Examples

[0507] Below are some examples of prompts to input to a generative AI model.

[0508] "Please make the sentence 'Please let me know the shipping status of the item' clearer."

[0509] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0510] Step 1:

[0511] The user enters the inquiry into a web form on the device. The input is text data, specifically a question such as "Please tell me the shipping status of my item." The user finishes entering the content into the web form as the output of this step.

[0512] Step 2:

[0513] The user clicks the "Get revision proposal" button on the device. This action causes the device to send the query to the server using JavaScript. The input data is the question entered by the user, and the output data is the request data in JSON format sent to the server.

[0514] Step 3:

[0515] The server receives the query sent from the terminal. It receives the question sent in JSON format as input. Next, it converts the question into tokens using a tokenization processing means. This converts the question into a format that can be input to the generative model. The server outputs the query in token format.

[0516] Step 4:

[0517] The server invokes the generative model and provides the tokenized question as input. The generative model generates clear refinement suggestions based on this input. For example, in response to the input "Please tell me the shipping status of my item," the model outputs the refinement suggestion "Please tell me in detail about the shipping status of the item I have currently ordered."

[0518] Step 5:

[0519] The server converts the generated revisions into JSON format and sends them back to the terminal. At this time, the input is the revisions output by the generative model, and the output is JSON format data.

[0520] Step 6:

[0521] The terminal receives the JSON data returned from the server. The input is the JSON data containing the revision suggestions returned from the server. The terminal parses this data and converts it into a format suitable for display on a web page. For example, JavaScript can be used to display the revision suggestions in HTML elements.

[0522] Step 7:

[0523] The terminal displays the converted revised version to the user. The user checks the displayed version and revises the question if necessary. The input is the converted revised version, and the output at this step is a clear query for the user to confirm.

[0524] Through this series of processing steps, the system of the present invention can automatically refine the content of the user's inquiry and convert it into clearer and more specific content.

[0525] (Application example 1)

[0526] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0527] When users make inquiries to customer support on online shopping sites, the content of their inquiries is often unclear, resulting in delayed responses. Furthermore, users have difficulty formulating appropriate inquiries, which reduces the efficiency of inquiry processes. To solve this problem, a method for making inquiries clear and specific is needed.

[0528] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0529] In this invention, the server includes an input means for inputting a query, a revision means for revising the query received from the input means using a generative model, a display means for displaying the query revised by the revision means, and a transmission means for confirming the displayed revision plan and transmitting the query. This enables a user to easily create a clear and specific query and receive a prompt and appropriate response.

[0530] The "inquiry content" is a specific description of the question or request that the user wants customer support to resolve.

[0531] The "input means" is a device or software that provides an interface for a user to input inquiry content to the system.

[0532] A "generative model" is an algorithm or program based on natural language processing that converts input queries into clearer, more specific sentences.

[0533] The "refining means" is a device or software that automatically refines the inquiry content received from the input means using a generative model and converts it into more appropriate inquiry content.

[0534] The "display means" is a device or software for visually presenting the inquiry content revised by the revision means to the user.

[0535] "Transmission means" refers to a device or software for transmitting the user-confirmed revision proposal to customer support.

[0536] This invention relates to a system for improving the efficiency of user inquiries on online shopping sites. The system automatically refines the content of inquiries entered by users and converts them into clear and specific inquiries. A specific embodiment of this system is described in detail below.

[0537] System configuration

[0538] Server side

[0539] The server receives the query and generates refinement proposals using the generative model. Specifically, it uses the following hardware and software:

[0540] Hardware:

[0541] General server computer

[0542] software:

[0543] Natural Language Processing Generative Model: Transformers (Hugging Face)

[0544] Programming language: Python

[0545] Web server: Node.js

[0546] Database: MongoDB

[0547] The server first tokenizes the query received from the user and inputs it into a generative model. The generative model then uses natural language processing technology to generate clearer and more specific query sentences. For example, if a user inputs "Please tell me the shipping status of my item," the server inputs this sentence into the generative model and generates a refinement suggestion: "Please tell me in detail the shipping status of the item I currently ordered." The server then returns this refinement suggestion to the terminal in JSON format.

[0548] Terminal side

[0549] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. Specifically, the following hardware and software are used:

[0550] Hardware:

[0551] Smartphone

[0552] software:

[0553] Frontend: React Native

[0554] API communication: Axios

[0555] The device provides a user interface (UI) for entering inquiries using a web form. When the user clicks the "Get revision suggestions" button, a request is sent to the server using JavaScript in the browser. After the server returns the revision suggestions, the device displays them on the screen.

[0556] User side

[0557] The user operates the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the device, the user can refer to the suggestions to revise the inquiry and click the send button to send it to customer support.

[0558] Specific examples

[0559] For example, if a user enters "Please tell me the shipping status of the product" into the inquiry form and clicks the "Get revision proposal" button, the following process will occur.

[0560] 1. User input

[0561] Input: "Please let me know the shipping status of the item."

[0562] 2. The server generates a revision plan

[0563] Suggested revision: "Please tell me the details of the shipping status of the item I ordered."

[0564] 3. The device will display the revised version.

[0565] Message: "Please tell me the details of the shipping status of the item I ordered."

[0566] 4. User reviews the revision and submits an inquiry

[0567] In this way, users can easily formulate clear and specific questions and receive prompt and appropriate responses.

[0568] This system will improve the efficiency of customer support operations and also improve the user experience.

[0569] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0570] Step 1:

[0571] The user enters a question into the inquiry form. The terminal acquires the input and temporarily stores it. The input in this case is "Please tell me the shipping status of the product."

[0572] Step 2:

[0573] The device sends the entered query to the server, packaging it in JSON format and sending it to the server's API endpoint using an HTTP POST request.

[0574] Step 3:

[0575] The server retrieves the received JSON data and extracts the query content. Here, it parses the data and converts it into a format that can be processed by the generative model.

[0576] Step 4:

[0577] The server inputs the query into the generative model and generates refinement suggestions. The generative model performs tokenization and appropriate embeddings to generate clear sentences. For example, the refinement suggestion generated is, "Please tell me the details of the shipping status of the item I've currently ordered."

[0578] Step 5:

[0579] The server packages the generated revision suggestions in JSON format and returns them to the terminal, sending JSON data containing the revision suggestions as an HTTP response.

[0580] Step 6:

[0581] The device receives the revision proposals from the server and displays them on the user interface by parsing the received JSON data, extracting the revision proposals, and displaying them on the screen.

[0582] Step 7:

[0583] The user checks the revised proposal displayed on the terminal, makes corrections as necessary, then confirms the revised inquiry content and clicks the send button to send it to customer support.

[0584] Step 8:

[0585] The device sends the inquiry confirmed by the user to the server again as the final inquiry, again in JSON format.

[0586] Step 9:

[0587] The server receives the final query and forwards it to the customer support team, where it is stored in a database and a response is initiated.

[0588] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0589] This invention is a system that combines a series of processes for inputting, revising, and displaying query content with an emotion engine that recognizes and analyzes user emotions. The system aims to generate more appropriate query sentences by automatically revising the query content entered by the user using a generative model and further analyzing the user's emotions.

[0590] System configuration

[0591] Server side

[0592] The server receives the query and generates refinement suggestions using a generative model. It also has the function of analyzing the emotions contained in the user's query using an emotion engine. The generative model generates text to clarify the question based on natural language processing.

[0593] For example, if a user enters "My product delivery is late, please tell me what's going on," the server inputs this question into the generative model and generates a refinement suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why the delivery is late." The emotion engine also analyzes the user's emotions from this question and recognizes emotions such as "dissatisfaction" and "irritation." Based on this information, the refinement suggestion is adjusted to generate a query that is more suited to the user.

[0594] The server hosts pre-trained models, e.g., using the Transformers library, as generative models, and also uses pre-trained emotion recognition models as emotion engines.

[0595] Terminal side

[0596] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[0597] For example, if a user inputs "The delivery of my product is late, please let me know what's going on" and clicks the "Get revision suggestions" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late," the terminal will display this revision suggestion on the web page.

[0598] User side

[0599] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0600] Specific examples

[0601] Specific examples are shown below.

[0602] The user enters "The delivery of my product is late, so please let me know what's going on." into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed." The emotion engine also analyzes emotions such as "dissatisfaction" and "irritation" from the question. Based on this information, the server adjusts the revision suggestion and generates a revision suggestion that takes the user's emotions into consideration, such as "I'm sorry to have kept you waiting. Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and revise the original question.

[0603] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[0604] The processing flow will be explained below.

[0605] Step 1:

[0606] The user inputs a question into the inquiry form.

[0607] The user types "My item is taking a long time to ship, please let me know what's going on" into a text area on a web form.

[0608] Step 2:

[0609] The user clicks the "Get Revision Suggestions" button.

[0610] The user clicks a button to request revision of the inquiry.

[0611] Step 3:

[0612] The terminal obtains the user's input.

[0613] The terminal uses JavaScript to obtain the question entered in the text area, "The shipment of the product is late, please let me know what's going on."

[0614] Code such as const questionText = document.getElementById("question").value; is executed.

[0615] Step 4:

[0616] The terminal uses the emotion engine to send a request to the server to recognize the user's emotion.

[0617] The device creates a request to send the question to the emotion engine for emotion analysis.

[0618] Code such as const emotionResponse = await fetch('http: / / server_address / emotion', {...}); is executed.

[0619] Step 5:

[0620] The server receives the query.

[0621] The server receives the POST request from the user terminal and analyzes the question.

[0622] Step 6:

[0623] The server uses an emotion engine to analyze the user's emotions.

[0624] The server inputs the question into an emotion engine and obtains emotional information such as "dissatisfaction" and "irritation."

[0625] Step 7:

[0626] The server uses the generative model to refine the text.

[0627] The server uses an AI generation model to tokenize the question, input it into the model, and generate a refined sentence.

[0628] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[0629] Step 8:

[0630] The server decodes the generated revisions and converts them into text.

[0631] The server decodes the generated token and converts it into human-readable text.

[0632] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[0633] Step 9:

[0634] The server takes into account the user's emotional information and adjusts the revision proposal.

[0635] Based on the results of the emotion engine, the server adds emotional considerations to the revised proposal, for example adding phrases such as "Sorry for keeping you waiting."

[0636] Step 10:

[0637] The server returns a response including a revision proposal to the terminal.

[0638] The server includes the adjusted revision proposal in a JSON format response and sends it back to the terminal.

[0639] Step 11:

[0640] The terminal receives the response from the server.

[0641] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[0642] Code such as const data = await response.json(); is executed.

[0643] Step 12:

[0644] The terminal displays the revised draft on the screen.

[0645] The terminal displays the received revision proposal in a specified HTML element on a web page.

[0646] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[0647] Step 13:

[0648] The user checks the revised draft.

[0649] The user views the revised question displayed on the screen, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know the reason for the delay in shipping." and compares it with the original question.

[0650] Step 14:

[0651] The user can modify the question as needed.

[0652] The user uses the revised suggestions as a reference and modifies the original question to something like, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[0653] These steps allow users to create clear, specific inquiries that are sensitive to emotions, improving work efficiency.

[0654] Example 2

[0655] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0656] In conventional inquiry systems, the inquiries entered by users are not always clear, or do not reflect the user's feelings, resulting in inappropriate inquiries. This results in poor efficiency in responding to inquiries and low user satisfaction. In particular, mechanical revisions that ignore the user's feelings can lead to poor communication quality and even cause problems. Therefore, there is a need for a system that can make the inquiries entered by users clearer and more appropriate, and that can respond in a way that takes the user's feelings into consideration.

[0657] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0658] In this invention, the server includes an input means for inputting the content of a query, a refining means for refining the received content of the query using a generative model, a display means for displaying the refined content of the query, an emotion recognition means for analyzing emotions from the generated content of the query, and an adjustment means for adjusting the content of the query based on the analyzed emotions. This makes it possible to make the content of the user's query clear and appropriate, and to generate an optimal query sentence that takes the user's emotions into consideration.

[0659] A "query" is the text a user enters to request specific information.

[0660] "Input means" refers to an interface or device that a user uses to input the contents of a query.

[0661] A "generative model" is an algorithm or software that uses natural language processing techniques to automatically refine input text.

[0662] The "elaboration means" refers to a means for modifying the input query content to make it clearer and more appropriate using a generative model.

[0663] "Display means" refers to an interface or device for visually presenting the refined inquiry content to the user.

[0664] "Emotion recognition means" refers to an algorithm or software that analyzes and recognizes a user's emotions from the content of the input inquiry.

[0665] The "adjustment means" refers to a means for optimizing the inquiry content based on the emotion information analyzed by the emotion recognition means.

[0666] "Interactive means" refers to an input interface or device that allows the user to review and modify the final query statement generated.

[0667] "Embedding techniques" refer to techniques used to make generative models correspond to specific knowledge and vocabulary.

[0668] The present invention is a system that combines a series of processes for inputting, revising, and displaying the content of a query with an emotion engine that recognizes and analyzes the user's emotions. Specific embodiments for carrying out the present invention will be described below.

[0669] Server side

[0670] The server receives the query content, generates refinement proposals using the generative model, and analyzes the emotions contained in the user's query content using an emotion engine.

[0671] The software used is the Transformers library, which is based on natural language processing technology, for the generative model. A pre-trained generative AI model is hosted on a server and inputs the query content sent by the user.

[0672] As a concrete example, if a user inputs "The delivery of my product is late, please let me know what's going on," the server will run this question through the generative model and generate a refinement suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late."

[0673] The server then uses an emotion recognition engine to recognize the emotions contained in the user's query. The emotion engine uses a pre-trained emotion recognition model that analyzes emotions from text and identifies emotions such as "frustration" or "irritation."

[0674] Based on this information, the server generates a final query, such as "I'm sorry for the wait. Please let me know the status of the delivery of my order. I'd also like to know why the delivery is delayed."

[0675] Terminal side

[0676] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. A web form is used as the specific user interface.

[0677] This involves the user entering their query in a text field and clicking the "Get Revision Suggestions" button. The device then uses JavaScript in the browser to send a request to the server. Specifically, it uses an Ajax request to send data asynchronously.

[0678] Once the response from the server is received, the device displays the revised version on the web page. For example, if the server returns a message like, "Sorry for the wait. Please let me know the status of the delivery of the item I ordered. I would also like to know why the delivery is delayed," this message will be displayed on the screen.

[0679] User side

[0680] The user enters a question into the inquiry form on the device and clicks the "Get revision proposal" button. This operation sends the inquiry to the server.

[0681] When the server returns the revised query, the content is displayed on the terminal. The user can review it and make corrections as necessary, which allows the user to create a clearer and more appropriate query.

[0682] Specific examples

[0683] Specific examples are shown below.

[0684] 1. The user enters "The product delivery is late, please let me know what's going on" into the inquiry form and clicks the "Get revision suggestions" button.

[0685] 2. The device sends this input to the server.

[0686] 3. The server uses the generative model to generate a refinement suggestion such as, "Please tell me the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[0687] 4. The emotion engine analyzes the input for emotions such as "dissatisfaction" or "irritation." Based on this information, it adjusts the suggested revision and generates a message like, "Sorry for the wait. Please let me know the status of the delivery of my order. I would also like to know why the delivery is delayed."

[0688] 5. The final query is sent back to the terminal, which displays it on the screen.

[0689] 6. The user checks the displayed revision suggestions and corrects the original question.

[0690] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[0691] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0692] Step 1: User enters inquiry

[0693] The user accesses a web form on the device and enters the inquiry in the text field. Specifically, the user opens the form in a browser and enters the prompt sentence, "The delivery of my product is late. Please let me know what's going on."

[0694] Input: Inquiry details

[0695] Output: The inquiry content is saved on the terminal.

[0696] Step 2: The device sends the input to the server

[0697] When the user clicks the "Get Revision Suggestions" button, the device uses JavaScript to send the query to the server. This operation is performed asynchronously using an Ajax request with the HTTP POST method.

[0698] Input: Inquiry details

[0699] Output: The query is sent to the server

[0700] Step 3: The server inputs the query into the generative model

[0701] The server passes the query received from the device through a generative model, which uses the Transformers library, a natural language processing technology, to generate clear refinement suggestions based on the query.

[0702] For example, the server converts an inquiry received such as "The shipment of the product is late, please tell me what's going on" into a refined proposal such as "Please tell me the shipping status of the product I ordered. I would also like to know the reason for the delay in shipping."

[0703] Input: Inquiry details

[0704] Data processing: Using generative AI models to refine queries

[0705] Output: Elaboration plan

[0706] Step 4: The server analyzes the emotion using the emotion recognition method.

[0707] The server inputs the generated revision plan into an emotion recognition means for analyzing the user's emotion, which uses an emotion recognition model to extract emotion information from the revision plan.

[0708] For example, emotions such as "dissatisfaction" and "irritation" are recognized as analysis results.

[0709] Input: Revision

[0710] Data Computing: Analyzing emotions using emotion recognition models

[0711] Output: Emotional information

[0712] Step 5: The server generates the final query based on the emotion information.

[0713] The server combines the generated refinement proposals with the emotion information to generate a final query that takes the user's emotions into consideration.

[0714] For example, you might say, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[0715] Input: Revision plan, emotional information

[0716] Data processing: Adjusting query sentences based on emotional information

[0717] Output: Final query

[0718] Step 6: The server sends the final query back to the terminal.

[0719] After the final query has been generated, the server sends it back to the terminal, which receives it and prepares it for display to the user.

[0720] Input: Final query

[0721] Output: The final query is sent to the terminal

[0722] Step 7: The terminal displays the final query

[0723] The final query returned by the server is displayed on a web page on the device, which uses HTML and JavaScript to display the text in a user-friendly way.

[0724] For example, the browser screen might display, "We apologize for the wait. Please let us know the shipping status of the item you ordered. We would also like to know the reason for the delay in shipping."

[0725] Input: Final query

[0726] Output: The query is displayed on the screen.

[0727] Step 8: User confirms and modifies the query

[0728] The user can review the final query displayed and make corrections as necessary, allowing the user to create a clear and appropriate query and proceed to final submission.

[0729] For example, the user confirms or corrects by saying, "Please let me know the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[0730] Input: Final query

[0731] Output: Modified query

[0732] (Application example 2)

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

[0734] When responding to inquiries in physical stores, it can be difficult for customers to enter appropriate and clear inquiry content. It is also necessary to properly recognize customer emotions and respond accordingly, but current systems are not sufficient in this regard. In particular, responding without understanding customer emotions carries the risk of lowering the quality of service. Facing these challenges, there is a need for the development of a system that allows customers to make inquiries in stores effectively and with consideration for their emotions.

[0735] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0736] In this invention, the server includes an input means for inputting an inquiry, a refining means for refining the inquiry received from the input means using a generative model, a display means for displaying the inquiry refined by the refining means, an emotion recognition means for analyzing the emotion of the inquiry, and an optimization means for optimizing the inquiry based on the emotion analyzed by the emotion recognition means. This makes it possible to clarify the inquiry input by a customer and to provide an optimized inquiry sentence that takes emotion into consideration.

[0737] "Input means" refers to a device or interface that allows a user to input inquiry details.

[0738] "Refining means" refers to a function that uses a generative model to revise the query entered by the user into a clearer and more specific sentence.

[0739] "Display means" refers to a device or interface for visually presenting the refined query content or optimized query sentence to the user.

[0740] "Emotion recognition means" refers to a function that analyzes emotions from the inquiry content entered by the user and identifies those emotions.

[0741] The "optimization means" refers to a function that appropriately adjusts the content of the inquiry based on the emotions analyzed by the emotion recognition means, and generates an inquiry statement that takes the user's emotions into consideration.

[0742] "Generative model" refers to a machine learning model that uses natural language processing techniques to generate and modify text.

[0743] This invention is a system that automatically refines customer inquiries and analyzes user sentiment to improve customer support in brick-and-mortar stores. The system consists of three main components: a server, a terminal, and a user.

[0744] Server side

[0745] The server receives the query content and processes it using a generative model and emotion recognition means.

[0746] 1. Generative Model

[0747] Hardware used: High-performance server computer

[0748] Software used: Natural language processing model using the Transformers library

[0749] Data computation: The received query content is input into the generative model to generate clear refinement suggestions.

[0750] 2. Emotion recognition means

[0751] Hardware used: High-performance server computer

[0752] Software used: Sentiment analysis model

[0753] Data calculation: Analyze emotions from inquiries and identify feelings such as "dissatisfaction" or "irritation."

[0754] 3. Optimization Methods

[0755] An optimized query is generated based on the refinement suggestions and the sentiment analysis results.

[0756] As a concrete example, if a user inputs "The delivery of my item is late, please let me know what's going on," the server inputs this into the generative model. The generative model generates a refinement suggestion such as "Please tell me the delivery status of the item I ordered. I would also like to know why the delivery is delayed." At the same time, the sentiment analysis model identifies emotions such as "dissatisfaction" and "irritation." Through optimization, the final sentence generated is "I'm sorry to have kept you waiting. Please let me know the delivery status of the item I ordered. I would also like to know why the delivery is delayed."

[0757] Terminal side

[0758] The terminal has the role of transmitting the inquiry content entered by the user to the server and displaying the response from the server.

[0759] 1. Interface

[0760] Hardware used: Smartphone, tablet

[0761] Software used: Web application (HTML, JavaScript)

[0762] Data Entry: User enters their query into the text field and presses the "Get Revisions" button

[0763] 2. Data transmission and reception

[0764] Software used: Data communication using AJAX

[0765] Data calculation: Sends the input query to the server and receives refinement suggestions and sentiment analysis results.

[0766] As a concrete example, a user inputs "The delivery of my product is late, please let me know what's going on," and clicks the "Get revision proposal" button. The device sends this content to the server, which returns a revision proposal saying, "I'm sorry to have kept you waiting. Please let me know the delivery status of the product I ordered. I'd also like to know why the delivery is late."

[0767] User side

[0768] The user is responsible for inputting and modifying the query using a terminal and receiving the optimized query statement from the server.

[0769] 1. Input

[0770] The user enters their inquiry into the text field.

[0771] 2. Correction and confirmation

[0772] Check the optimized query and modify it if necessary

[0773] For example, a user can input "The shipment of my item is delayed, so please let me know what's going on." and then review the generated sentence "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know why the shipment is delayed." and correct the original inquiry.

[0774] The present invention improves the efficiency of responding to inquiries in physical stores and also realizes optimal responses that take into consideration the feelings of users.

[0775] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0776] Step 1:

[0777] The user inputs the inquiry into the terminal interface. Specifically, the user inputs the inquiry into the text field and clicks the "Get revision proposal" button. The input data is the inquiry in text format.

[0778] Step 2:

[0779] The terminal sends the entered query content to the server. In this process, the input content is sent to the server using asynchronous communication using AJAX. The input data is the text content entered by the user and is sent to the server as output.

[0780] Step 3:

[0781] The server inputs the query received into a generative model to generate a refinement proposal. Specifically, it uses the Transformers library to refine the query into a clearer and more specific sentence. The input data is the user's query, and the output data is the text refined by the generative model.

[0782] Step 4:

[0783] The server inputs the refined query content into the emotion recognition means to analyze the emotion. An emotion analysis model is used to extract emotions such as "dissatisfaction" or "irritation" from the text. The input data is the refined query sentence, and the output data is the analyzed emotion information.

[0784] Step 5:

[0785] The server optimizes the query content based on the emotion recognition results. Specifically, it adjusts the wording to be appropriate, taking into account the emotion information. The input data is the emotion analysis results and the refined query, and the output data is the optimized query.

[0786] Step 6:

[0787] The server sends the optimized query statement to the terminal. In this process, the generated optimized statement is returned to the terminal. The input data is the optimized query statement, and it is sent to the terminal as output data.

[0788] Step 7:

[0789] The terminal displays the optimized query received by the terminal to the user. Specifically, the query received from the server is displayed on the interface so that the user can confirm and modify it. The input data is the optimized query, and the output data is the text content displayed to the user.

[0790] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0791] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0792] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[0793] [Third embodiment]

[0794] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[0795] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0796] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0797] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[0798] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0799] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0800] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0801] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0802] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0803] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0804] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0805] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[0806] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically revise query content entered by a user using a generative model and convert it into a clear query content.

[0807] System configuration

[0808] Server side

[0809] The server receives the query and generates a refinement proposal using a generative model. The generative model generates text to clarify the question based on natural language processing. For example, if a user enters "Please tell me the shipping status of my item," the server inputs this question into the generative model and generates a refinement proposal such as "Please tell me in detail the shipping status of the item I currently ordered."

[0810] The server hosts a pre-trained model, for example using the Transformers library, as a generative model. The server tokenizes the acquired question, inputs it into the generative model, and outputs the revised sentence. The server then returns this revised sentence in JSON format to the user's device.

[0811] Terminal side

[0812] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[0813] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision suggestion" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me in detail about the shipping status of the product I have currently ordered," the terminal will display this revision suggestion on the web page.

[0814] User side

[0815] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0816] Specific examples

[0817] Specific examples are shown below.

[0818] The user enters "Please tell me the shipping status of my item" into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the item I have currently ordered," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and modify the original question.

[0819] As a result, the system of the present invention can improve the clarity of the inquiry content and improve the efficiency of the business process.

[0820] The processing flow will be explained below.

[0821] Step 1:

[0822] The user inputs a question into the inquiry form.

[0823] The user enters "Please tell me the shipping status of my item" into the text area of ​​the web form.

[0824] Step 2:

[0825] The user clicks the "Get Revision Suggestions" button.

[0826] The user clicks a button to request revision of the inquiry.

[0827] Step 3:

[0828] The terminal obtains the user's input.

[0829] The terminal uses JavaScript to obtain the question "Please tell me the shipping status of the product" entered in the text area.

[0830] Code such as const questionText = document.getElementById("question").value; is executed.

[0831] Step 4:

[0832] The terminal sends a request to the server.

[0833] The device uses the fetch API to send a POST request containing the question to the server.

[0834] Code such as const response = await fetch('http: / / server_address / suggestions', {...}); is executed.

[0835] Step 5:

[0836] The server receives the request.

[0837] The server receives a POST request from the user terminal. The question is included in the request body.

[0838] Step 6:

[0839] The server uses the generative model to refine the text.

[0840] The server uses an AI model to tokenize the question, input it into the model, and generate a refined sentence.

[0841] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[0842] Step 7:

[0843] The server decodes the generated revisions and converts them into text.

[0844] The server decodes the generated token and converts it into human-readable text.

[0845] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[0846] Step 8:

[0847] The server returns a response including a revision proposal to the terminal.

[0848] The server includes the generated revision proposal in a response in JSON format and sends it back to the terminal.

[0849] Step 9:

[0850] The terminal receives the response from the server.

[0851] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[0852] Code such as const data = await response.json(); is executed.

[0853] Step 10:

[0854] The terminal displays the revised draft on the screen.

[0855] The terminal displays the received revision proposal in a specified HTML element on a web page.

[0856] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[0857] Step 11:

[0858] The user checks the revised draft.

[0859] The user views the revised question displayed on the screen, "Please tell me in detail about the shipping status of the item I have currently ordered," and compares it with the original question.

[0860] Step 12:

[0861] The user can modify the question as needed.

[0862] The user uses the revised suggestions as a reference and modifies the original question to something like, "Please tell me in detail about the shipping status of the item I have currently ordered."

[0863] These steps allow users to formulate clear and specific inquiries, improving work efficiency.

[0864] Example 1

[0865] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0866] In today's information society, users make numerous inquiries, but these are often expressed vaguely or unclearly. This makes it difficult for the recipient (e.g., a company or support staff) to accurately understand the intent, and it can take time to respond. Furthermore, if the inquiry content is unclear, the quality of the response will decline, which in turn reduces user satisfaction. To solve this problem, a system is needed that can automatically refine the inquiry content and convert it into clear and specific content.

[0867] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0868] In this invention, the server includes an input means for inputting an inquiry, a refinement means for refining the inquiry received from the input means using a generative model, a processing means for tokenizing the inquiry for input to the generative model, a return means for returning the inquiry refined by the refinement means in JSON format, and a display means for displaying the inquiry refined by the refinement means. This automatically converts the inquiry input by the user into clear and specific content, making it possible to improve the efficiency of inquiry operations and the quality of responses.

[0869] "Inquiry content" refers to a question or request for information that a user sends to the system through an input means.

[0870] "Input means" refers to a device or software that provides an interface for a user to input inquiry details.

[0871] A "generative model" refers to an algorithm or machine learning model that automatically generates refinement suggestions based on the specified inquiry content.

[0872] "Refining means" refers to the process or function that uses a generative model to refine the content of a received query for clarity.

[0873] "Processing means" refers to the process or function that converts the query content into an appropriate form before inputting it into the generative model.

[0874] "JSON format" stands for JavaScript Object Notation and refers to a lightweight data exchange format for structuring data.

[0875] The "returning means" refers to a process or function for returning the generated revision proposal to the user or terminal.

[0876] "Display means" refers to a device or software for displaying the refined inquiry content on a terminal in a format that can be confirmed by the user.

[0877] "Embedding technology" refers to a data representation technique that enables generative models to correspond to specific knowledge and terminology.

[0878] "Interactive means" refers to a device or software that provides an interface to allow a user to select or modify a refinement.

[0879] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically refine query content entered by a user using a generative model and convert it into an unambiguous query content.

[0880] Server side

[0881] The server receives the query and generates a refinement proposal using a generative model. This generative model uses natural language processing. Specifically, the server processes the query using a tokenization processing means and generates a refinement proposal.

[0882] The server hosts a pre-trained generative model using the Transformers library, a powerful tool implemented in Python and specialized for natural language processing. The server takes the user's question, tokenizes it, and feeds it into the generative model. The generative model generates clearer text and returns its refinements in JSON format.

[0883] For example, if a user inputs an inquiry such as "Please tell me the shipping status of the product," the server inputs this information into the generative model and generates a refinement proposal such as "Please tell me in detail about the shipping status of the product I have currently ordered."

[0884] Terminal side

[0885] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying revision suggestions. Specifically, the terminal has a web form where the user enters their inquiry. When the user clicks the "Get revision suggestions" button to submit the input, a request is sent to the server using JavaScript.

[0886] The terminal receives the response from the server and displays the received revision suggestions on the screen, allowing the user to check the revision suggestions and revise the original question.

[0887] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision proposal" button, the terminal will send this inquiry to the server. If the server returns a revision proposal saying "Please tell me the details of the shipping status of the product I have currently ordered," the terminal will display this revision proposal on the web page.

[0888] User side

[0889] The user begins operating the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the server displays the revision suggestions on the device, the user can use these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0890] Specific examples

[0891] A concrete example is given below. A user enters "Please tell me the shipping status of my product" into an inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the product I have currently ordered," and sends it back to the device. The device displays this revision suggestion on a web page. The user checks the displayed revision suggestion and modifies the original question.

[0892] Prompt Sentence Examples

[0893] Below are some examples of prompts to input to a generative AI model.

[0894] "Please make the sentence 'Please let me know the shipping status of the item' clearer."

[0895] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0896] Step 1:

[0897] The user enters the inquiry into a web form on the device. The input is text data, specifically a question such as "Please tell me the shipping status of my item." The user finishes entering the content into the web form as the output of this step.

[0898] Step 2:

[0899] The user clicks the "Get revision proposal" button on the device. This action causes the device to send the query to the server using JavaScript. The input data is the question entered by the user, and the output data is the request data in JSON format sent to the server.

[0900] Step 3:

[0901] The server receives the query sent from the terminal. It receives the question sent in JSON format as input. Next, it converts the question into tokens using a tokenization processing means. This converts the question into a format that can be input to the generative model. The server outputs the query in token format.

[0902] Step 4:

[0903] The server invokes the generative model and provides the tokenized question as input. The generative model generates clear refinement suggestions based on this input. For example, in response to the input "Please tell me the shipping status of my item," the model outputs the refinement suggestion "Please tell me in detail about the shipping status of the item I have currently ordered."

[0904] Step 5:

[0905] The server converts the generated revisions into JSON format and sends them back to the terminal. At this time, the input is the revisions output by the generative model, and the output is JSON format data.

[0906] Step 6:

[0907] The terminal receives the JSON data returned from the server. The input is the JSON data containing the revision suggestions returned from the server. The terminal parses this data and converts it into a format suitable for display on a web page. For example, JavaScript can be used to display the revision suggestions in HTML elements.

[0908] Step 7:

[0909] The terminal displays the converted revised version to the user. The user checks the displayed version and revises the question if necessary. The input is the converted revised version, and the output at this step is a clear query for the user to confirm.

[0910] Through this series of processing steps, the system of the present invention can automatically refine the content of the user's inquiry and convert it into clearer and more specific content.

[0911] (Application example 1)

[0912] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[0913] When users make inquiries to customer support on online shopping sites, the content of their inquiries is often unclear, resulting in delayed responses. Furthermore, users have difficulty formulating appropriate inquiries, which reduces the efficiency of inquiry processes. To solve this problem, a method for making inquiries clear and specific is needed.

[0914] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0915] In this invention, the server includes an input means for inputting a query, a revision means for revising the query received from the input means using a generative model, a display means for displaying the query revised by the revision means, and a transmission means for confirming the displayed revision plan and transmitting the query. This enables a user to easily create a clear and specific query and receive a prompt and appropriate response.

[0916] The "inquiry content" is a specific description of the question or request that the user wants customer support to resolve.

[0917] The "input means" is a device or software that provides an interface for a user to input inquiry content to the system.

[0918] A "generative model" is an algorithm or program based on natural language processing that converts input queries into clearer, more specific sentences.

[0919] The "refining means" is a device or software that automatically refines the inquiry content received from the input means using a generative model and converts it into more appropriate inquiry content.

[0920] The "display means" is a device or software for visually presenting the inquiry content revised by the revision means to the user.

[0921] "Transmission means" refers to a device or software for transmitting the user-confirmed revision proposal to customer support.

[0922] This invention relates to a system for improving the efficiency of user inquiries on online shopping sites. The system automatically refines the content of inquiries entered by users and converts them into clear and specific inquiries. A specific embodiment of this system is described in detail below.

[0923] System configuration

[0924] Server side

[0925] The server receives the query and generates refinement proposals using the generative model. Specifically, it uses the following hardware and software:

[0926] Hardware:

[0927] General server computer

[0928] software:

[0929] Natural Language Processing Generative Model: Transformers (Hugging Face)

[0930] Programming language: Python

[0931] Web server: Node.js

[0932] Database: MongoDB

[0933] The server first tokenizes the query received from the user and inputs it into a generative model. The generative model then uses natural language processing technology to generate clearer and more specific query sentences. For example, if a user inputs "Please tell me the shipping status of my item," the server inputs this sentence into the generative model and generates a refinement suggestion: "Please tell me in detail the shipping status of the item I currently ordered." The server then returns this refinement suggestion to the terminal in JSON format.

[0934] Terminal side

[0935] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. Specifically, the following hardware and software are used:

[0936] Hardware:

[0937] Smartphone

[0938] software:

[0939] Frontend: React Native

[0940] API communication: Axios

[0941] The device provides a user interface (UI) for entering inquiries using a web form. When the user clicks the "Get revision suggestions" button, a request is sent to the server using JavaScript in the browser. After the server returns the revision suggestions, the device displays them on the screen.

[0942] User side

[0943] The user operates the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the device, the user can refer to the suggestions to revise the inquiry and click the send button to send it to customer support.

[0944] Specific examples

[0945] For example, if a user enters "Please tell me the shipping status of the product" into the inquiry form and clicks the "Get revision proposal" button, the following process will occur.

[0946] 1. User input

[0947] Input: "Please let me know the shipping status of the item."

[0948] 2. The server generates a revision plan

[0949] Suggested revision: "Please tell me the details of the shipping status of the item I ordered."

[0950] 3. The device will display the revised version.

[0951] Message: "Please tell me the details of the shipping status of the item I ordered."

[0952] 4. User reviews the revision and submits an inquiry

[0953] In this way, users can easily formulate clear and specific questions and receive prompt and appropriate responses.

[0954] This system will improve the efficiency of customer support operations and also improve the user experience.

[0955] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0956] Step 1:

[0957] The user enters a question into the inquiry form. The terminal acquires the input and temporarily stores it. The input in this case is "Please tell me the shipping status of the product."

[0958] Step 2:

[0959] The device sends the entered query to the server, packaging it in JSON format and sending it to the server's API endpoint using an HTTP POST request.

[0960] Step 3:

[0961] The server retrieves the received JSON data and extracts the query content. Here, it parses the data and converts it into a format that can be processed by the generative model.

[0962] Step 4:

[0963] The server inputs the query into the generative model and generates refinement suggestions. The generative model performs tokenization and appropriate embeddings to generate clear sentences. For example, the refinement suggestion generated is, "Please tell me the details of the shipping status of the item I've currently ordered."

[0964] Step 5:

[0965] The server packages the generated revision suggestions in JSON format and returns them to the terminal, sending JSON data containing the revision suggestions as an HTTP response.

[0966] Step 6:

[0967] The device receives the revision proposals from the server and displays them on the user interface by parsing the received JSON data, extracting the revision proposals, and displaying them on the screen.

[0968] Step 7:

[0969] The user checks the revised proposal displayed on the terminal, makes corrections as necessary, then confirms the revised inquiry content and clicks the send button to send it to customer support.

[0970] Step 8:

[0971] The device sends the inquiry confirmed by the user to the server again as the final inquiry, again in JSON format.

[0972] Step 9:

[0973] The server receives the final query and forwards it to the customer support team, where it is stored in a database and a response is initiated.

[0974] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0975] This invention is a system that combines a series of processes for inputting, revising, and displaying query content with an emotion engine that recognizes and analyzes user emotions. The system aims to generate more appropriate query sentences by automatically revising the query content entered by the user using a generative model and further analyzing the user's emotions.

[0976] System configuration

[0977] Server side

[0978] The server receives the query and generates refinement suggestions using a generative model. It also has the function of analyzing the emotions contained in the user's query using an emotion engine. The generative model generates text to clarify the question based on natural language processing.

[0979] For example, if a user enters "My product delivery is late, please tell me what's going on," the server inputs this question into the generative model and generates a refinement suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why the delivery is late." The emotion engine also analyzes the user's emotions from this question and recognizes emotions such as "dissatisfaction" and "irritation." Based on this information, the refinement suggestion is adjusted to generate a query that is more suited to the user.

[0980] The server hosts pre-trained models, e.g., using the Transformers library, as generative models, and also uses pre-trained emotion recognition models as emotion engines.

[0981] Terminal side

[0982] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[0983] For example, if a user inputs "The delivery of my product is late, please let me know what's going on" and clicks the "Get revision suggestions" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late," the terminal will display this revision suggestion on the web page.

[0984] User side

[0985] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[0986] Specific examples

[0987] Specific examples are shown below.

[0988] The user enters "The delivery of my product is late, so please let me know what's going on." into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed." The emotion engine also analyzes emotions such as "dissatisfaction" and "irritation" from the question. Based on this information, the server adjusts the revision suggestion and generates a revision suggestion that takes the user's emotions into consideration, such as "I'm sorry to have kept you waiting. Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and revise the original question.

[0989] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[0990] The processing flow will be explained below.

[0991] Step 1:

[0992] The user inputs a question into the inquiry form.

[0993] The user types "My item is taking a long time to ship, please let me know what's going on" into a text area on a web form.

[0994] Step 2:

[0995] The user clicks the "Get Revision Suggestions" button.

[0996] The user clicks a button to request revision of the inquiry.

[0997] Step 3:

[0998] The terminal obtains the user's input.

[0999] The terminal uses JavaScript to obtain the question entered in the text area, "The shipment of the product is late, please let me know what's going on."

[1000] Code such as const questionText = document.getElementById("question").value; is executed.

[1001] Step 4:

[1002] The terminal uses the emotion engine to send a request to the server to recognize the user's emotion.

[1003] The device creates a request to send the question to the emotion engine for emotion analysis.

[1004] Code such as const emotionResponse = await fetch('http: / / server_address / emotion', {...}); is executed.

[1005] Step 5:

[1006] The server receives the query.

[1007] The server receives the POST request from the user terminal and analyzes the question.

[1008] Step 6:

[1009] The server uses an emotion engine to analyze the user's emotions.

[1010] The server inputs the question into an emotion engine and obtains emotional information such as "dissatisfaction" and "irritation."

[1011] Step 7:

[1012] The server uses the generative model to refine the text.

[1013] The server uses an AI generation model to tokenize the question, input it into the model, and generate a refined sentence.

[1014] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[1015] Step 8:

[1016] The server decodes the generated revisions and converts them into text.

[1017] The server decodes the generated token and converts it into human-readable text.

[1018] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[1019] Step 9:

[1020] The server takes into account the user's emotional information and adjusts the revision proposal.

[1021] Based on the results of the emotion engine, the server adds emotional considerations to the revised proposal, for example adding phrases such as "Sorry for keeping you waiting."

[1022] Step 10:

[1023] The server returns a response including a revision proposal to the terminal.

[1024] The server includes the adjusted revision proposal in a JSON format response and sends it back to the terminal.

[1025] Step 11:

[1026] The terminal receives the response from the server.

[1027] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[1028] Code such as const data = await response.json(); is executed.

[1029] Step 12:

[1030] The terminal displays the revised draft on the screen.

[1031] The terminal displays the received revision proposal in a specified HTML element on a web page.

[1032] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[1033] Step 13:

[1034] The user checks the revised draft.

[1035] The user views the revised question displayed on the screen, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know the reason for the delay in shipping." and compares it with the original question.

[1036] Step 14:

[1037] The user can modify the question as needed.

[1038] The user uses the revised suggestions as a reference and modifies the original question to something like, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[1039] These steps allow users to create clear, specific inquiries that are sensitive to emotions, improving work efficiency.

[1040] Example 2

[1041] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1042] In conventional inquiry systems, the inquiries entered by users are not always clear, or do not reflect the user's feelings, resulting in inappropriate inquiries. This results in poor efficiency in responding to inquiries and low user satisfaction. In particular, mechanical revisions that ignore the user's feelings can lead to poor communication quality and even cause problems. Therefore, there is a need for a system that can make the inquiries entered by users clearer and more appropriate, and that can respond in a way that takes the user's feelings into consideration.

[1043] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1044] In this invention, the server includes an input means for inputting the content of a query, a refining means for refining the received content of the query using a generative model, a display means for displaying the refined content of the query, an emotion recognition means for analyzing emotions from the generated content of the query, and an adjustment means for adjusting the content of the query based on the analyzed emotions. This makes it possible to make the content of the user's query clear and appropriate, and to generate an optimal query sentence that takes the user's emotions into consideration.

[1045] A "query" is the text a user enters to request specific information.

[1046] "Input means" refers to an interface or device that a user uses to input the contents of a query.

[1047] A "generative model" is an algorithm or software that uses natural language processing techniques to automatically refine input text.

[1048] The "elaboration means" refers to a means for modifying the input query content to make it clearer and more appropriate using a generative model.

[1049] "Display means" refers to an interface or device for visually presenting the refined inquiry content to the user.

[1050] "Emotion recognition means" refers to an algorithm or software that analyzes and recognizes a user's emotions from the content of the input inquiry.

[1051] The "adjustment means" refers to a means for optimizing the inquiry content based on the emotion information analyzed by the emotion recognition means.

[1052] "Interactive means" refers to an input interface or device that allows the user to review and modify the final query statement generated.

[1053] "Embedding techniques" refer to techniques used to make generative models correspond to specific knowledge and vocabulary.

[1054] The present invention is a system that combines a series of processes for inputting, revising, and displaying the content of a query with an emotion engine that recognizes and analyzes the user's emotions. Specific embodiments for carrying out the present invention will be described below.

[1055] Server side

[1056] The server receives the query content, generates refinement proposals using the generative model, and analyzes the emotions contained in the user's query content using an emotion engine.

[1057] The software used is the Transformers library, which is based on natural language processing technology, for the generative model. A pre-trained generative AI model is hosted on a server and inputs the query content sent by the user.

[1058] As a concrete example, if a user inputs "The delivery of my product is late, please let me know what's going on," the server will run this question through the generative model and generate a refinement suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late."

[1059] The server then uses an emotion recognition engine to recognize the emotions contained in the user's query. The emotion engine uses a pre-trained emotion recognition model that analyzes emotions from text and identifies emotions such as "frustration" or "irritation."

[1060] Based on this information, the server generates a final query, such as "I'm sorry for the wait. Please let me know the status of the delivery of my order. I'd also like to know why the delivery is delayed."

[1061] Terminal side

[1062] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. A web form is used as the specific user interface.

[1063] This involves the user entering their query in a text field and clicking the "Get Revision Suggestions" button. The device then uses JavaScript in the browser to send a request to the server. Specifically, it uses an Ajax request to send data asynchronously.

[1064] Once the response from the server is received, the device displays the revised version on the web page. For example, if the server returns a message like, "Sorry for the wait. Please let me know the status of the delivery of the item I ordered. I would also like to know why the delivery is delayed," this message will be displayed on the screen.

[1065] User side

[1066] The user enters a question into the inquiry form on the device and clicks the "Get revision proposal" button. This operation sends the inquiry to the server.

[1067] When the server returns the revised query, the content is displayed on the terminal. The user can review it and make corrections as necessary, which allows the user to create a clearer and more appropriate query.

[1068] Specific examples

[1069] Specific examples are shown below.

[1070] 1. The user enters "The product delivery is late, please let me know what's going on" into the inquiry form and clicks the "Get revision suggestions" button.

[1071] 2. The device sends this input to the server.

[1072] 3. The server uses the generative model to generate a refinement suggestion such as, "Please tell me the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[1073] 4. The emotion engine analyzes the input for emotions such as "dissatisfaction" or "irritation." Based on this information, it adjusts the suggested revision and generates a message like, "Sorry for the wait. Please let me know the status of the delivery of my order. I would also like to know why the delivery is delayed."

[1074] 5. The final query is sent back to the terminal, which displays it on the screen.

[1075] 6. The user checks the displayed revision suggestions and corrects the original question.

[1076] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

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

[1078] Step 1: User enters inquiry

[1079] The user accesses a web form on the device and enters the inquiry in the text field. Specifically, the user opens the form in a browser and enters the prompt sentence, "The delivery of my product is late. Please let me know what's going on."

[1080] Input: Inquiry details

[1081] Output: The inquiry content is saved on the terminal.

[1082] Step 2: The device sends the input to the server

[1083] When the user clicks the "Get Revision Suggestions" button, the device uses JavaScript to send the query to the server. This operation is performed asynchronously using an Ajax request with the HTTP POST method.

[1084] Input: Inquiry details

[1085] Output: The query is sent to the server

[1086] Step 3: The server inputs the query into the generative model

[1087] The server passes the query received from the device through a generative model, which uses the Transformers library, a natural language processing technology, to generate clear refinement suggestions based on the query.

[1088] For example, the server converts an inquiry received such as "The shipment of the product is late, please tell me what's going on" into a refined proposal such as "Please tell me the shipping status of the product I ordered. I would also like to know the reason for the delay in shipping."

[1089] Input: Inquiry details

[1090] Data processing: Using generative AI models to refine queries

[1091] Output: Elaboration plan

[1092] Step 4: The server analyzes the emotion using the emotion recognition method.

[1093] The server inputs the generated revision plan into an emotion recognition means for analyzing the user's emotion, which uses an emotion recognition model to extract emotion information from the revision plan.

[1094] For example, emotions such as "dissatisfaction" and "irritation" are recognized as analysis results.

[1095] Input: Revision

[1096] Data Computing: Analyzing emotions using emotion recognition models

[1097] Output: Emotional information

[1098] Step 5: The server generates the final query based on the emotion information.

[1099] The server combines the generated refinement proposals with the emotion information to generate a final query that takes the user's emotions into consideration.

[1100] For example, you might say, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[1101] Input: Revision plan, emotional information

[1102] Data processing: Adjusting query sentences based on emotional information

[1103] Output: Final query

[1104] Step 6: The server sends the final query back to the terminal.

[1105] After the final query has been generated, the server sends it back to the terminal, which receives it and prepares it for display to the user.

[1106] Input: Final query

[1107] Output: The final query is sent to the terminal

[1108] Step 7: The terminal displays the final query

[1109] The final query returned by the server is displayed on a web page on the device, which uses HTML and JavaScript to display the text in a user-friendly way.

[1110] For example, the browser screen might display, "We apologize for the wait. Please let us know the shipping status of the item you ordered. We would also like to know the reason for the delay in shipping."

[1111] Input: Final query

[1112] Output: The query is displayed on the screen.

[1113] Step 8: User confirms and modifies the query

[1114] The user can review the final query displayed and make corrections as necessary, allowing the user to create a clear and appropriate query and proceed to final submission.

[1115] For example, the user confirms or corrects by saying, "Please let me know the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[1116] Input: Final query

[1117] Output: Modified query

[1118] (Application example 2)

[1119] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1120] When responding to inquiries in physical stores, it can be difficult for customers to enter appropriate and clear inquiry content. It is also necessary to properly recognize customer emotions and respond accordingly, but current systems are not sufficient in this regard. In particular, responding without understanding customer emotions carries the risk of lowering the quality of service. Facing these challenges, there is a need for the development of a system that allows customers to make inquiries in stores effectively and with consideration for their emotions.

[1121] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1122] In this invention, the server includes an input means for inputting an inquiry, a refining means for refining the inquiry received from the input means using a generative model, a display means for displaying the inquiry refined by the refining means, an emotion recognition means for analyzing the emotion of the inquiry, and an optimization means for optimizing the inquiry based on the emotion analyzed by the emotion recognition means. This makes it possible to clarify the inquiry input by a customer and to provide an optimized inquiry sentence that takes emotion into consideration.

[1123] "Input means" refers to a device or interface that allows a user to input inquiry details.

[1124] "Refining means" refers to a function that uses a generative model to revise the query entered by the user into a clearer and more specific sentence.

[1125] "Display means" refers to a device or interface for visually presenting the refined query content or optimized query sentence to the user.

[1126] "Emotion recognition means" refers to a function that analyzes emotions from the inquiry content entered by the user and identifies those emotions.

[1127] The "optimization means" refers to a function that appropriately adjusts the content of the inquiry based on the emotions analyzed by the emotion recognition means, and generates an inquiry statement that takes the user's emotions into consideration.

[1128] "Generative model" refers to a machine learning model that uses natural language processing techniques to generate and modify text.

[1129] This invention is a system that automatically refines customer inquiries and analyzes user sentiment to improve customer support in brick-and-mortar stores. The system consists of three main components: a server, a terminal, and a user.

[1130] Server side

[1131] The server receives the query content and processes it using a generative model and emotion recognition means.

[1132] 1. Generative Model

[1133] Hardware used: High-performance server computer

[1134] Software used: Natural language processing model using the Transformers library

[1135] Data computation: The received query content is input into the generative model to generate clear refinement suggestions.

[1136] 2. Emotion recognition means

[1137] Hardware used: High-performance server computer

[1138] Software used: Sentiment analysis model

[1139] Data calculation: Analyze emotions from inquiries and identify feelings such as "dissatisfaction" or "irritation."

[1140] 3. Optimization Methods

[1141] An optimized query is generated based on the refinement suggestions and the sentiment analysis results.

[1142] As a concrete example, if a user inputs "The delivery of my item is late, please let me know what's going on," the server inputs this into the generative model. The generative model generates a refinement suggestion such as "Please tell me the delivery status of the item I ordered. I would also like to know why the delivery is delayed." At the same time, the sentiment analysis model identifies emotions such as "dissatisfaction" and "irritation." Through optimization, the final sentence generated is "I'm sorry to have kept you waiting. Please let me know the delivery status of the item I ordered. I would also like to know why the delivery is delayed."

[1143] Terminal side

[1144] The terminal has the role of transmitting the inquiry content entered by the user to the server and displaying the response from the server.

[1145] 1. Interface

[1146] Hardware used: Smartphone, tablet

[1147] Software used: Web application (HTML, JavaScript)

[1148] Data Entry: User enters their query into the text field and presses the "Get Revisions" button

[1149] 2. Data transmission and reception

[1150] Software used: Data communication using AJAX

[1151] Data calculation: Sends the input query to the server and receives refinement suggestions and sentiment analysis results.

[1152] As a concrete example, a user inputs "The delivery of my product is late, please let me know what's going on," and clicks the "Get revision proposal" button. The device sends this content to the server, which returns a revision proposal saying, "I'm sorry to have kept you waiting. Please let me know the delivery status of the product I ordered. I'd also like to know why the delivery is late."

[1153] User side

[1154] The user is responsible for inputting and modifying the query using a terminal and receiving the optimized query statement from the server.

[1155] 1. Input

[1156] The user enters their inquiry into the text field.

[1157] 2. Correction and confirmation

[1158] Check the optimized query and modify it if necessary

[1159] For example, a user can input "The shipment of my item is delayed, so please let me know what's going on." and then review the generated sentence "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know why the shipment is delayed." and correct the original inquiry.

[1160] The present invention improves the efficiency of responding to inquiries in physical stores and also realizes optimal responses that take into consideration the feelings of users.

[1161] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1162] Step 1:

[1163] The user inputs the inquiry into the terminal interface. Specifically, the user inputs the inquiry into the text field and clicks the "Get revision proposal" button. The input data is the inquiry in text format.

[1164] Step 2:

[1165] The terminal sends the entered query content to the server. In this process, the input content is sent to the server using asynchronous communication using AJAX. The input data is the text content entered by the user and is sent to the server as output.

[1166] Step 3:

[1167] The server inputs the query received into a generative model to generate a refinement proposal. Specifically, it uses the Transformers library to refine the query into a clearer and more specific sentence. The input data is the user's query, and the output data is the text refined by the generative model.

[1168] Step 4:

[1169] The server inputs the refined query content into the emotion recognition means to analyze the emotion. An emotion analysis model is used to extract emotions such as "dissatisfaction" or "irritation" from the text. The input data is the refined query sentence, and the output data is the analyzed emotion information.

[1170] Step 5:

[1171] The server optimizes the query content based on the emotion recognition results. Specifically, it adjusts the wording to be appropriate, taking into account the emotion information. The input data is the emotion analysis results and the refined query, and the output data is the optimized query.

[1172] Step 6:

[1173] The server sends the optimized query statement to the terminal. In this process, the generated optimized statement is returned to the terminal. The input data is the optimized query statement, and it is sent to the terminal as output data.

[1174] Step 7:

[1175] The terminal displays the optimized query received by the terminal to the user. Specifically, the query received from the server is displayed on the interface so that the user can confirm and modify it. The input data is the optimized query, and the output data is the text content displayed to the user.

[1176] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1177] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1178] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1179] [Fourth embodiment]

[1180] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1181] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1182] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1183] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1184] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1185] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1186] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1187] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1188] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1189] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1190] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1191] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1192] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1193] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically revise query content entered by a user using a generative model and convert it into a clear query content.

[1194] System configuration

[1195] Server side

[1196] The server receives the query and generates a refinement proposal using a generative model. The generative model generates text to clarify the question based on natural language processing. For example, if a user enters "Please tell me the shipping status of my item," the server inputs this question into the generative model and generates a refinement proposal such as "Please tell me in detail the shipping status of the item I currently ordered."

[1197] The server hosts a pre-trained model, for example using the Transformers library, as a generative model. The server tokenizes the acquired question, inputs it into the generative model, and outputs the revised sentence. The server then returns this revised sentence in JSON format to the user's device.

[1198] Terminal side

[1199] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[1200] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision suggestion" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me in detail about the shipping status of the product I have currently ordered," the terminal will display this revision suggestion on the web page.

[1201] User side

[1202] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[1203] Specific examples

[1204] Specific examples are shown below.

[1205] The user enters "Please tell me the shipping status of my item" into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the item I have currently ordered," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and modify the original question.

[1206] As a result, the system of the present invention can improve the clarity of the inquiry content and improve the efficiency of the business process.

[1207] The processing flow will be explained below.

[1208] Step 1:

[1209] The user inputs a question into the inquiry form.

[1210] The user enters "Please tell me the shipping status of my item" into the text area of ​​the web form.

[1211] Step 2:

[1212] The user clicks the "Get Revision Suggestions" button.

[1213] The user clicks a button to request revision of the inquiry.

[1214] Step 3:

[1215] The terminal obtains the user's input.

[1216] The terminal uses JavaScript to obtain the question "Please tell me the shipping status of the product" entered in the text area.

[1217] Code such as const questionText = document.getElementById("question").value; is executed.

[1218] Step 4:

[1219] The terminal sends a request to the server.

[1220] The device uses the fetch API to send a POST request containing the question to the server.

[1221] Code such as const response = await fetch('http: / / server_address / suggestions', {...}); is executed.

[1222] Step 5:

[1223] The server receives the request.

[1224] The server receives a POST request from the user terminal. The question is included in the request body.

[1225] Step 6:

[1226] The server uses the generative model to refine the text.

[1227] The server uses an AI model to tokenize the question, input it into the model, and generate a refined sentence.

[1228] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[1229] Step 7:

[1230] The server decodes the generated revisions and converts them into text.

[1231] The server decodes the generated token and converts it into human-readable text.

[1232] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[1233] Step 8:

[1234] The server returns a response including a revision proposal to the terminal.

[1235] The server includes the generated revision proposal in a response in JSON format and sends it back to the terminal.

[1236] Step 9:

[1237] The terminal receives the response from the server.

[1238] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[1239] Code such as const data = await response.json(); is executed.

[1240] Step 10:

[1241] The terminal displays the revised draft on the screen.

[1242] The terminal displays the received revision proposal in a specified HTML element on a web page.

[1243] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[1244] Step 11:

[1245] The user checks the revised draft.

[1246] The user views the revised question displayed on the screen, "Please tell me in detail about the shipping status of the item I have currently ordered," and compares it with the original question.

[1247] Step 12:

[1248] The user can modify the question as needed.

[1249] The user uses the revised suggestions as a reference and modifies the original question to something like, "Please tell me in detail about the shipping status of the item I have currently ordered."

[1250] These steps allow users to formulate clear and specific inquiries, improving work efficiency.

[1251] Example 1

[1252] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1253] In today's information society, users make numerous inquiries, but these are often expressed vaguely or unclearly. This makes it difficult for the recipient (e.g., a company or support staff) to accurately understand the intent, and it can take time to respond. Furthermore, if the inquiry content is unclear, the quality of the response will decline, which in turn reduces user satisfaction. To solve this problem, a system is needed that can automatically refine the inquiry content and convert it into clear and specific content.

[1254] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1255] In this invention, the server includes an input means for inputting an inquiry, a refinement means for refining the inquiry received from the input means using a generative model, a processing means for tokenizing the inquiry for input to the generative model, a return means for returning the inquiry refined by the refinement means in JSON format, and a display means for displaying the inquiry refined by the refinement means. This automatically converts the inquiry input by the user into clear and specific content, making it possible to improve the efficiency of inquiry operations and the quality of responses.

[1256] "Inquiry content" refers to a question or request for information that a user sends to the system through an input means.

[1257] "Input means" refers to a device or software that provides an interface for a user to input inquiry details.

[1258] A "generative model" refers to an algorithm or machine learning model that automatically generates refinement suggestions based on the specified inquiry content.

[1259] "Refining means" refers to the process or function that uses a generative model to refine the content of a received query for clarity.

[1260] "Processing means" refers to the process or function that converts the query content into an appropriate form before inputting it into the generative model.

[1261] "JSON format" stands for JavaScript Object Notation and refers to a lightweight data exchange format for structuring data.

[1262] The "returning means" refers to a process or function for returning the generated revision proposal to the user or terminal.

[1263] "Display means" refers to a device or software for displaying the refined inquiry content on a terminal in a format that can be confirmed by the user.

[1264] "Embedding technology" refers to a data representation technique that enables generative models to correspond to specific knowledge and terminology.

[1265] "Interactive means" refers to a device or software that provides an interface to allow a user to select or modify a refinement.

[1266] The present invention provides a system for implementing a series of processes for inputting, revising, and displaying query content. The system aims to automatically refine query content entered by a user using a generative model and convert it into an unambiguous query content.

[1267] Server side

[1268] The server receives the query and generates a refinement proposal using a generative model. This generative model uses natural language processing. Specifically, the server processes the query using a tokenization processing means and generates a refinement proposal.

[1269] The server hosts a pre-trained generative model using the Transformers library, a powerful tool implemented in Python and specialized for natural language processing. The server takes the user's question, tokenizes it, and feeds it into the generative model. The generative model generates clearer text and returns its refinements in JSON format.

[1270] For example, if a user inputs an inquiry such as "Please tell me the shipping status of the product," the server inputs this information into the generative model and generates a refinement proposal such as "Please tell me in detail about the shipping status of the product I have currently ordered."

[1271] Terminal side

[1272] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying revision suggestions. Specifically, the terminal has a web form where the user enters their inquiry. When the user clicks the "Get revision suggestions" button to submit the input, a request is sent to the server using JavaScript.

[1273] The terminal receives the response from the server and displays the received revision suggestions on the screen, allowing the user to check the revision suggestions and revise the original question.

[1274] For example, if a user inputs "Please tell me the shipping status of my product" and clicks the "Get revision proposal" button, the terminal will send this inquiry to the server. If the server returns a revision proposal saying "Please tell me the details of the shipping status of the product I have currently ordered," the terminal will display this revision proposal on the web page.

[1275] User side

[1276] The user begins operating the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the server displays the revision suggestions on the device, the user can use these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[1277] Specific examples

[1278] A concrete example is given below. A user enters "Please tell me the shipping status of my product" into an inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion, "Please tell me in detail about the shipping status of the product I have currently ordered," and sends it back to the device. The device displays this revision suggestion on a web page. The user checks the displayed revision suggestion and modifies the original question.

[1279] Prompt Sentence Examples

[1280] Below are some examples of prompts to input to a generative AI model.

[1281] "Please make the sentence 'Please let me know the shipping status of the item' clearer."

[1282] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1283] Step 1:

[1284] The user enters the inquiry into a web form on the device. The input is text data, specifically a question such as "Please tell me the shipping status of my item." The user finishes entering the content into the web form as the output of this step.

[1285] Step 2:

[1286] The user clicks the "Get revision proposal" button on the device. This action causes the device to send the query to the server using JavaScript. The input data is the question entered by the user, and the output data is the request data in JSON format sent to the server.

[1287] Step 3:

[1288] The server receives the query sent from the terminal. It receives the question sent in JSON format as input. Next, it converts the question into tokens using a tokenization processing means. This converts the question into a format that can be input to the generative model. The server outputs the query in token format.

[1289] Step 4:

[1290] The server invokes the generative model and provides the tokenized question as input. The generative model generates clear refinement suggestions based on this input. For example, in response to the input "Please tell me the shipping status of my item," the model outputs the refinement suggestion "Please tell me in detail about the shipping status of the item I have currently ordered."

[1291] Step 5:

[1292] The server converts the generated revisions into JSON format and sends them back to the terminal. At this time, the input is the revisions output by the generative model, and the output is JSON format data.

[1293] Step 6:

[1294] The terminal receives the JSON data returned from the server. The input is the JSON data containing the revision suggestions returned from the server. The terminal parses this data and converts it into a format suitable for display on a web page. For example, JavaScript can be used to display the revision suggestions in HTML elements.

[1295] Step 7:

[1296] The terminal displays the converted revised version to the user. The user checks the displayed version and revises the question if necessary. The input is the converted revised version, and the output at this step is a clear query for the user to confirm.

[1297] Through this series of processing steps, the system of the present invention can automatically refine the content of the user's inquiry and convert it into clearer and more specific content.

[1298] (Application example 1)

[1299] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1300] When users make inquiries to customer support on online shopping sites, the content of their inquiries is often unclear, resulting in delayed responses. Furthermore, users have difficulty formulating appropriate inquiries, which reduces the efficiency of inquiry processes. To solve this problem, a method for making inquiries clear and specific is needed.

[1301] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1302] In this invention, the server includes an input means for inputting a query, a revision means for revising the query received from the input means using a generative model, a display means for displaying the query revised by the revision means, and a transmission means for confirming the displayed revision plan and transmitting the query. This enables a user to easily create a clear and specific query and receive a prompt and appropriate response.

[1303] The "inquiry content" is a specific description of the question or request that the user wants customer support to resolve.

[1304] The "input means" is a device or software that provides an interface for a user to input inquiry content to the system.

[1305] A "generative model" is an algorithm or program based on natural language processing that converts input queries into clearer, more specific sentences.

[1306] The "refining means" is a device or software that automatically refines the inquiry content received from the input means using a generative model and converts it into more appropriate inquiry content.

[1307] The "display means" is a device or software for visually presenting the inquiry content revised by the revision means to the user.

[1308] "Transmission means" refers to a device or software for transmitting the user-confirmed revision proposal to customer support.

[1309] This invention relates to a system for improving the efficiency of user inquiries on online shopping sites. The system automatically refines the content of inquiries entered by users and converts them into clear and specific inquiries. A specific embodiment of this system is described in detail below.

[1310] System configuration

[1311] Server side

[1312] The server receives the query and generates refinement proposals using the generative model. Specifically, it uses the following hardware and software:

[1313] Hardware:

[1314] General server computer

[1315] software:

[1316] Natural Language Processing Generative Model: Transformers (Hugging Face)

[1317] Programming language: Python

[1318] Web server: Node.js

[1319] Database: MongoDB

[1320] The server first tokenizes the query received from the user and inputs it into a generative model. The generative model then uses natural language processing technology to generate clearer and more specific query sentences. For example, if a user inputs "Please tell me the shipping status of my item," the server inputs this sentence into the generative model and generates a refinement suggestion: "Please tell me in detail the shipping status of the item I currently ordered." The server then returns this refinement suggestion to the terminal in JSON format.

[1321] Terminal side

[1322] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. Specifically, the following hardware and software are used:

[1323] Hardware:

[1324] Smartphone

[1325] software:

[1326] Frontend: React Native

[1327] API communication: Axios

[1328] The device provides a user interface (UI) for entering inquiries using a web form. When the user clicks the "Get revision suggestions" button, a request is sent to the server using JavaScript in the browser. After the server returns the revision suggestions, the device displays them on the screen.

[1329] User side

[1330] The user operates the system by entering a question into the inquiry form on the device and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the device, the user can refer to the suggestions to revise the inquiry and click the send button to send it to customer support.

[1331] Specific examples

[1332] For example, if a user enters "Please tell me the shipping status of the product" into the inquiry form and clicks the "Get revision proposal" button, the following process will occur.

[1333] 1. User input

[1334] Input: "Please let me know the shipping status of the item."

[1335] 2. The server generates a revision plan

[1336] Suggested revision: "Please tell me the details of the shipping status of the item I ordered."

[1337] 3. The device will display the revised version.

[1338] Message: "Please tell me the details of the shipping status of the item I ordered."

[1339] 4. User reviews the revision and submits an inquiry

[1340] In this way, users can easily formulate clear and specific questions and receive prompt and appropriate responses.

[1341] This system will improve the efficiency of customer support operations and also improve the user experience.

[1342] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1343] Step 1:

[1344] The user enters a question into the inquiry form. The terminal acquires the input and temporarily stores it. The input in this case is "Please tell me the shipping status of the product."

[1345] Step 2:

[1346] The device sends the entered query to the server, packaging it in JSON format and sending it to the server's API endpoint using an HTTP POST request.

[1347] Step 3:

[1348] The server retrieves the received JSON data and extracts the query content. Here, it parses the data and converts it into a format that can be processed by the generative model.

[1349] Step 4:

[1350] The server inputs the query into the generative model and generates refinement suggestions. The generative model performs tokenization and appropriate embeddings to generate clear sentences. For example, the refinement suggestion generated is, "Please tell me the details of the shipping status of the item I've currently ordered."

[1351] Step 5:

[1352] The server packages the generated revision suggestions in JSON format and returns them to the terminal, sending JSON data containing the revision suggestions as an HTTP response.

[1353] Step 6:

[1354] The device receives the revision proposals from the server and displays them on the user interface by parsing the received JSON data, extracting the revision proposals, and displaying them on the screen.

[1355] Step 7:

[1356] The user checks the revised proposal displayed on the terminal, makes corrections as necessary, then confirms the revised inquiry content and clicks the send button to send it to customer support.

[1357] Step 8:

[1358] The device sends the inquiry confirmed by the user to the server again as the final inquiry, again in JSON format.

[1359] Step 9:

[1360] The server receives the final query and forwards it to the customer support team, where it is stored in a database and a response is initiated.

[1361] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1362] This invention is a system that combines a series of processes for inputting, revising, and displaying query content with an emotion engine that recognizes and analyzes user emotions. The system aims to generate more appropriate query sentences by automatically revising the query content entered by the user using a generative model and further analyzing the user's emotions.

[1363] System configuration

[1364] Server side

[1365] The server receives the query and generates refinement suggestions using a generative model. It also has the function of analyzing the emotions contained in the user's query using an emotion engine. The generative model generates text to clarify the question based on natural language processing.

[1366] For example, if a user enters "My product delivery is late, please tell me what's going on," the server inputs this question into the generative model and generates a refinement suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why the delivery is late." The emotion engine also analyzes the user's emotions from this question and recognizes emotions such as "dissatisfaction" and "irritation." Based on this information, the refinement suggestion is adjusted to generate a query that is more suited to the user.

[1367] The server hosts pre-trained models, e.g., using the Transformers library, as generative models, and also uses pre-trained emotion recognition models as emotion engines.

[1368] Terminal side

[1369] The terminal is responsible for receiving input from the user, sending it to the server, and receiving and displaying the revision proposal. Specifically, it provides a UI for entering inquiry details using a web form, and when the "Get revision proposal" button is clicked, a request is sent to the server using JavaScript on the browser. The terminal receives the response from the server and displays the revision proposal on the screen.

[1370] For example, if a user inputs "The delivery of my product is late, please let me know what's going on" and clicks the "Get revision suggestions" button, the terminal will send the inquiry to the server. If the server returns a revision suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late," the terminal will display this revision suggestion on the web page.

[1371] User side

[1372] The user begins operating the system by entering a question into the inquiry form on the terminal and clicking the "Get revision suggestions" button. After the revision suggestions from the server are displayed on the terminal, the user can refer to these suggestions to revise the question. This allows the user to make clearer and more specific inquiries, improving the efficiency of inquiry work.

[1373] Specific examples

[1374] Specific examples are shown below.

[1375] The user enters "The delivery of my product is late, so please let me know what's going on." into the inquiry form and clicks the "Get revision suggestions" button. The device sends this question to the server. The server uses the generative model to generate a revision suggestion such as "Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed." The emotion engine also analyzes emotions such as "dissatisfaction" and "irritation" from the question. Based on this information, the server adjusts the revision suggestion and generates a revision suggestion that takes the user's emotions into consideration, such as "I'm sorry to have kept you waiting. Please tell me the shipping status of the product I ordered. I would also like to know why it's delayed," and sends it back to the device. The device displays this revision suggestion on the screen. The user can check the displayed revision suggestion and revise the original question.

[1376] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[1377] The processing flow will be explained below.

[1378] Step 1:

[1379] The user inputs a question into the inquiry form.

[1380] The user types "My item is taking a long time to ship, please let me know what's going on" into a text area on a web form.

[1381] Step 2:

[1382] The user clicks the "Get Revision Suggestions" button.

[1383] The user clicks a button to request revision of the inquiry.

[1384] Step 3:

[1385] The terminal obtains the user's input.

[1386] The terminal uses JavaScript to obtain the question entered in the text area, "The shipment of the product is late, please let me know what's going on."

[1387] Code such as const questionText = document.getElementById("question").value; is executed.

[1388] Step 4:

[1389] The terminal uses the emotion engine to send a request to the server to recognize the user's emotion.

[1390] The device creates a request to send the question to the emotion engine for emotion analysis.

[1391] Code such as const emotionResponse = await fetch('http: / / server_address / emotion', {...}); is executed.

[1392] Step 5:

[1393] The server receives the query.

[1394] The server receives the POST request from the user terminal and analyzes the question.

[1395] Step 6:

[1396] The server uses an emotion engine to analyze the user's emotions.

[1397] The server inputs the question into an emotion engine and obtains emotional information such as "dissatisfaction" and "irritation."

[1398] Step 7:

[1399] The server uses the generative model to refine the text.

[1400] The server uses an AI generation model to tokenize the question, input it into the model, and generate a refined sentence.

[1401] After inputs = tokenizer.encode("paraphrase: " + questionText, return_tensors="pt", max_length=512, truncation=True);, outputs = model.generate(inputs, max_length=512, num_return_sequences=1); is executed.

[1402] Step 8:

[1403] The server decodes the generated revisions and converts them into text.

[1404] The server decodes the generated token and converts it into human-readable text.

[1405] Code such as suggestion = tokenizer.decode(outputs[0], skip_special_tokens=True); is executed.

[1406] Step 9:

[1407] The server takes into account the user's emotional information and adjusts the revision proposal.

[1408] Based on the results of the emotion engine, the server adds emotional considerations to the revised proposal, for example adding phrases such as "Sorry for keeping you waiting."

[1409] Step 10:

[1410] The server returns a response including a revision proposal to the terminal.

[1411] The server includes the adjusted revision proposal in a JSON format response and sends it back to the terminal.

[1412] Step 11:

[1413] The terminal receives the response from the server.

[1414] The device uses the fetch API to receive the response from the server and parses the response body as JSON.

[1415] Code such as const data = await response.json(); is executed.

[1416] Step 12:

[1417] The terminal displays the revised draft on the screen.

[1418] The terminal displays the received revision proposal in a specified HTML element on a web page.

[1419] Code such as document.getElementById("suggestion").innerText = data.suggestion; is executed.

[1420] Step 13:

[1421] The user checks the revised draft.

[1422] The user views the revised question displayed on the screen, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know the reason for the delay in shipping." and compares it with the original question.

[1423] Step 14:

[1424] The user can modify the question as needed.

[1425] The user uses the revised suggestions as a reference and modifies the original question to something like, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[1426] These steps allow users to create clear, specific inquiries that are sensitive to emotions, improving work efficiency.

[1427] Example 2

[1428] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1429] In conventional inquiry systems, the inquiries entered by users are not always clear, or do not reflect the user's feelings, resulting in inappropriate inquiries. This results in poor efficiency in responding to inquiries and low user satisfaction. In particular, mechanical revisions that ignore the user's feelings can lead to poor communication quality and even cause problems. Therefore, there is a need for a system that can make the inquiries entered by users clearer and more appropriate, and that can respond in a way that takes the user's feelings into consideration.

[1430] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1431] In this invention, the server includes an input means for inputting the content of a query, a refining means for refining the received content of the query using a generative model, a display means for displaying the refined content of the query, an emotion recognition means for analyzing emotions from the generated content of the query, and an adjustment means for adjusting the content of the query based on the analyzed emotions. This makes it possible to make the content of the user's query clear and appropriate, and to generate an optimal query sentence that takes the user's emotions into consideration.

[1432] A "query" is the text a user enters to request specific information.

[1433] "Input means" refers to an interface or device that a user uses to input the contents of a query.

[1434] A "generative model" is an algorithm or software that uses natural language processing techniques to automatically refine input text.

[1435] The "elaboration means" refers to a means for modifying the input query content to make it clearer and more appropriate using a generative model.

[1436] "Display means" refers to an interface or device for visually presenting the refined inquiry content to the user.

[1437] "Emotion recognition means" refers to an algorithm or software that analyzes and recognizes a user's emotions from the content of the input inquiry.

[1438] The "adjustment means" refers to a means for optimizing the inquiry content based on the emotion information analyzed by the emotion recognition means.

[1439] "Interactive means" refers to an input interface or device that allows the user to review and modify the final query statement generated.

[1440] "Embedding techniques" refer to techniques used to make generative models correspond to specific knowledge and vocabulary.

[1441] The present invention is a system that combines a series of processes for inputting, revising, and displaying the content of a query with an emotion engine that recognizes and analyzes the user's emotions. Specific embodiments for carrying out the present invention will be described below.

[1442] Server side

[1443] The server receives the query content, generates refinement proposals using the generative model, and analyzes the emotions contained in the user's query content using an emotion engine.

[1444] The software used is the Transformers library, which is based on natural language processing technology, for the generative model. A pre-trained generative AI model is hosted on a server and inputs the query content sent by the user.

[1445] As a concrete example, if a user inputs "The delivery of my product is late, please let me know what's going on," the server will run this question through the generative model and generate a refinement suggestion such as "Please tell me the delivery status of the product I ordered. I would also like to know why the delivery is late."

[1446] The server then uses an emotion recognition engine to recognize the emotions contained in the user's query. The emotion engine uses a pre-trained emotion recognition model that analyzes emotions from text and identifies emotions such as "frustration" or "irritation."

[1447] Based on this information, the server generates a final query, such as "I'm sorry for the wait. Please let me know the status of the delivery of my order. I'd also like to know why the delivery is delayed."

[1448] Terminal side

[1449] The terminal receives input from the user, transmits it to the server, and receives and displays the revised draft. A web form is used as the specific user interface.

[1450] This involves the user entering their query in a text field and clicking the "Get Revision Suggestions" button. The device then uses JavaScript in the browser to send a request to the server. Specifically, it uses an Ajax request to send data asynchronously.

[1451] Once the response from the server is received, the device displays the revised version on the web page. For example, if the server returns a message like, "Sorry for the wait. Please let me know the status of the delivery of the item I ordered. I would also like to know why the delivery is delayed," this message will be displayed on the screen.

[1452] User side

[1453] The user enters a question into the inquiry form on the device and clicks the "Get revision proposal" button. This operation sends the inquiry to the server.

[1454] When the server returns the revised query, the content is displayed on the terminal. The user can review it and make corrections as necessary, which allows the user to create a clearer and more appropriate query.

[1455] Specific examples

[1456] Specific examples are shown below.

[1457] 1. The user enters "The product delivery is late, please let me know what's going on" into the inquiry form and clicks the "Get revision suggestions" button.

[1458] 2. The device sends this input to the server.

[1459] 3. The server uses the generative model to generate a refinement suggestion such as, "Please tell me the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[1460] 4. The emotion engine analyzes the input for emotions such as "dissatisfaction" or "irritation." Based on this information, it adjusts the suggested revision and generates a message like, "Sorry for the wait. Please let me know the status of the delivery of my order. I would also like to know why the delivery is delayed."

[1461] 5. The final query is sent back to the terminal, which displays it on the screen.

[1462] 6. The user checks the displayed revision suggestions and corrects the original question.

[1463] As a result, the system of the present invention can improve the clarity of the inquiry content and realize a response that takes into consideration the user's feelings, thereby improving the efficiency of business processes.

[1464] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1465] Step 1: User enters inquiry

[1466] The user accesses a web form on the device and enters the inquiry in the text field. Specifically, the user opens the form in a browser and enters the prompt sentence, "The delivery of my product is late. Please let me know what's going on."

[1467] Input: Inquiry details

[1468] Output: The inquiry content is saved on the terminal.

[1469] Step 2: The device sends the input to the server

[1470] When the user clicks the "Get Revision Suggestions" button, the device uses JavaScript to send the query to the server. This operation is performed asynchronously using an Ajax request with the HTTP POST method.

[1471] Input: Inquiry details

[1472] Output: The query is sent to the server

[1473] Step 3: The server inputs the query into the generative model

[1474] The server passes the query received from the device through a generative model, which uses the Transformers library, a natural language processing technology, to generate clear refinement suggestions based on the query.

[1475] For example, the server converts an inquiry received such as "The shipment of the product is late, please tell me what's going on" into a refined proposal such as "Please tell me the shipping status of the product I ordered. I would also like to know the reason for the delay in shipping."

[1476] Input: Inquiry details

[1477] Data processing: Using generative AI models to refine queries

[1478] Output: Elaboration plan

[1479] Step 4: The server analyzes the emotion using the emotion recognition method.

[1480] The server inputs the generated revision plan into an emotion recognition means for analyzing the user's emotion, which uses an emotion recognition model to extract emotion information from the revision plan.

[1481] For example, emotions such as "dissatisfaction" and "irritation" are recognized as analysis results.

[1482] Input: Revision

[1483] Data Computing: Analyzing emotions using emotion recognition models

[1484] Output: Emotional information

[1485] Step 5: The server generates the final query based on the emotion information.

[1486] The server combines the generated refinement proposals with the emotion information to generate a final query that takes the user's emotions into consideration.

[1487] For example, you might say, "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I'd also like to know why the shipment is delayed."

[1488] Input: Revision plan, emotional information

[1489] Data processing: Adjusting query sentences based on emotional information

[1490] Output: Final query

[1491] Step 6: The server sends the final query back to the terminal.

[1492] After the final query has been generated, the server sends it back to the terminal, which receives it and prepares it for display to the user.

[1493] Input: Final query

[1494] Output: The final query is sent to the terminal

[1495] Step 7: The terminal displays the final query

[1496] The final query returned by the server is displayed on a web page on the device, which uses HTML and JavaScript to display the text in a user-friendly way.

[1497] For example, the browser screen might display, "We apologize for the wait. Please let us know the shipping status of the item you ordered. We would also like to know the reason for the delay in shipping."

[1498] Input: Final query

[1499] Output: The query is displayed on the screen.

[1500] Step 8: User confirms and modifies the query

[1501] The user can review the final query displayed and make corrections as necessary, allowing the user to create a clear and appropriate query and proceed to final submission.

[1502] For example, the user confirms or corrects by saying, "Please let me know the shipping status of the item I ordered. I would also like to know why the shipping is delayed."

[1503] Input: Final query

[1504] Output: Modified query

[1505] (Application example 2)

[1506] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1507] When responding to inquiries in physical stores, it can be difficult for customers to enter appropriate and clear inquiry content. It is also necessary to properly recognize customer emotions and respond accordingly, but current systems are not sufficient in this regard. In particular, responding without understanding customer emotions carries the risk of lowering the quality of service. Facing these challenges, there is a need for the development of a system that allows customers to make inquiries in stores effectively and with consideration for their emotions.

[1508] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1509] In this invention, the server includes an input means for inputting an inquiry, a refining means for refining the inquiry received from the input means using a generative model, a display means for displaying the inquiry refined by the refining means, an emotion recognition means for analyzing the emotion of the inquiry, and an optimization means for optimizing the inquiry based on the emotion analyzed by the emotion recognition means. This makes it possible to clarify the inquiry input by a customer and to provide an optimized inquiry sentence that takes emotion into consideration.

[1510] "Input means" refers to a device or interface that allows a user to input inquiry details.

[1511] "Refining means" refers to a function that uses a generative model to revise the query entered by the user into a clearer and more specific sentence.

[1512] "Display means" refers to a device or interface for visually presenting the refined query content or optimized query sentence to the user.

[1513] "Emotion recognition means" refers to a function that analyzes emotions from the inquiry content entered by the user and identifies those emotions.

[1514] The "optimization means" refers to a function that appropriately adjusts the content of the inquiry based on the emotions analyzed by the emotion recognition means, and generates an inquiry statement that takes the user's emotions into consideration.

[1515] "Generative model" refers to a machine learning model that uses natural language processing techniques to generate and modify text.

[1516] This invention is a system that automatically refines customer inquiries and analyzes user sentiment to improve customer support in brick-and-mortar stores. The system consists of three main components: a server, a terminal, and a user.

[1517] Server side

[1518] The server receives the query content and processes it using a generative model and emotion recognition means.

[1519] 1. Generative Model

[1520] Hardware used: High-performance server computer

[1521] Software used: Natural language processing model using the Transformers library

[1522] Data computation: The received query content is input into the generative model to generate clear refinement suggestions.

[1523] 2. Emotion recognition means

[1524] Hardware used: High-performance server computer

[1525] Software used: Sentiment analysis model

[1526] Data calculation: Analyze emotions from inquiries and identify feelings such as "dissatisfaction" or "irritation."

[1527] 3. Optimization Methods

[1528] An optimized query is generated based on the refinement suggestions and the sentiment analysis results.

[1529] As a concrete example, if a user inputs "The delivery of my item is late, please let me know what's going on," the server inputs this into the generative model. The generative model generates a refinement suggestion such as "Please tell me the delivery status of the item I ordered. I would also like to know why the delivery is delayed." At the same time, the sentiment analysis model identifies emotions such as "dissatisfaction" and "irritation." Through optimization, the final sentence generated is "I'm sorry to have kept you waiting. Please let me know the delivery status of the item I ordered. I would also like to know why the delivery is delayed."

[1530] Terminal side

[1531] The terminal has the role of transmitting the inquiry content entered by the user to the server and displaying the response from the server.

[1532] 1. Interface

[1533] Hardware used: Smartphone, tablet

[1534] Software used: Web application (HTML, JavaScript)

[1535] Data Entry: User enters their query into the text field and presses the "Get Revisions" button

[1536] 2. Data transmission and reception

[1537] Software used: Data communication using AJAX

[1538] Data calculation: Sends the input query to the server and receives refinement suggestions and sentiment analysis results.

[1539] As a concrete example, a user inputs "The delivery of my product is late, please let me know what's going on," and clicks the "Get revision proposal" button. The device sends this content to the server, which returns a revision proposal saying, "I'm sorry to have kept you waiting. Please let me know the delivery status of the product I ordered. I'd also like to know why the delivery is late."

[1540] User side

[1541] The user is responsible for inputting and modifying the query using a terminal and receiving the optimized query statement from the server.

[1542] 1. Input

[1543] The user enters their inquiry into the text field.

[1544] 2. Correction and confirmation

[1545] Check the optimized query and modify it if necessary

[1546] For example, a user can input "The shipment of my item is delayed, so please let me know what's going on." and then review the generated sentence "I'm sorry to have kept you waiting. Please let me know the shipping status of the item I ordered. I would also like to know why the shipment is delayed." and correct the original inquiry.

[1547] The present invention improves the efficiency of responding to inquiries in physical stores and also realizes optimal responses that take into consideration the feelings of users.

[1548] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1549] Step 1:

[1550] The user inputs the inquiry into the terminal interface. Specifically, the user inputs the inquiry into the text field and clicks the "Get revision proposal" button. The input data is the inquiry in text format.

[1551] Step 2:

[1552] The terminal sends the entered query content to the server. In this process, the input content is sent to the server using asynchronous communication using AJAX. The input data is the text content entered by the user and is sent to the server as output.

[1553] Step 3:

[1554] The server inputs the query received into a generative model to generate a refinement proposal. Specifically, it uses the Transformers library to refine the query into a clearer and more specific sentence. The input data is the user's query, and the output data is the text refined by the generative model.

[1555] Step 4:

[1556] The server inputs the refined query content into the emotion recognition means to analyze the emotion. An emotion analysis model is used to extract emotions such as "dissatisfaction" or "irritation" from the text. The input data is the refined query sentence, and the output data is the analyzed emotion information.

[1557] Step 5:

[1558] The server optimizes the query content based on the emotion recognition results. Specifically, it adjusts the wording to be appropriate, taking into account the emotion information. The input data is the emotion analysis results and the refined query, and the output data is the optimized query.

[1559] Step 6:

[1560] The server sends the optimized query statement to the terminal. In this process, the generated optimized statement is returned to the terminal. The input data is the optimized query statement, and it is sent to the terminal as output data.

[1561] Step 7:

[1562] The terminal displays the optimized query received by the terminal to the user. Specifically, the query received from the server is displayed on the interface so that the user can confirm and modify it. The input data is the optimized query, and the output data is the text content displayed to the user.

[1563] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The microphone 238 acquires voice indicating a user input regarding the result of the specific processing. The control unit 46A transmits voice data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the voice data.

[1564] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1565] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the robot 414.

[1566] The emotion identification model 59 as an emotion engine may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to an emotion map (see FIG. 9), which is a specific mapping. Similarly, the emotion identification model 59 may determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1567] FIG. 9 is a diagram illustrating an emotion map 400 on which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. Emotions closer to the center of the concentric circles are more primitive. Emotions representing states and actions arising from a state of mind are arranged on the outer edges of the concentric circles. The concept of emotion includes both affect and mental states. Emotions generally generated from reactions occurring in the brain are arranged on the left side of the concentric circles. Emotions generally induced by situational judgment are arranged on the right side of the concentric circles. Emotions generally generated from reactions occurring in the brain and induced by situational judgment are arranged on the upper and lower sides of the concentric circles. Furthermore, the emotion of "pleasure" is arranged on the upper side of the concentric circles, and the emotion of "discomfort" is arranged on the lower side. In this way, in the emotion map 400, multiple emotions are mapped based on the structure by which emotions are generated, and emotions that tend to occur simultaneously are mapped close to each other.

[1568] These emotions are distributed in the 3 o'clock direction on emotion map 400, and typically fluctuate between relief and anxiety. In the right half of emotion map 400, situational awareness dominates over internal sensations, resulting in a sense of calm.

[1569] The inside of emotion map 400 represents what is going on in the mind, and the outside of emotion map 400 represents behavior, so the further you go outside emotion map 400, the more visible the emotions become (the more they are expressed in behavior).

[1570] Human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. Emotions can also be created for robots, automobiles, and motorcycles, based on various balances, such as posture and remaining battery life. When these balances deviate from the ideal, a state of discomfort is indicated, and when they approach the ideal, a state of pleasure is indicated. An emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on Voice Emotion Recognition and Emotional Brain Physiological Signal Analysis Systems, Tokushima University, Doctoral Dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map lists emotions belonging to the "reaction" domain, where sensation is dominant. The right half of the emotion map lists emotions belonging to the "situation" domain, where situational awareness is dominant.

[1571] The emotion map defines two emotions that promote learning. One is a negative emotion on the situation side, around the middle of "repentance" or "reflection." In other words, this occurs when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is a positive emotion on the response side, around "desire." In other words, this occurs when the robot experiences positive feelings such as "I want more" or "I want to know more."

[1572] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​indicating each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple pieces of training data that are combinations of user input and emotion values ​​indicating each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions that are located close to each other have similar values, as in the emotion map 900 shown in FIG. 10. FIG. 10 shows an example in which multiple emotions, "relieved," "calm," and "reassuring," have similar emotion values.

[1573] The system according to the present disclosure has been described above mainly with respect to the functions of the data processing device 12, but the system according to the present disclosure is not necessarily implemented on a server. The system according to the present disclosure may be implemented as a general information processing system. The present disclosure may be implemented, for example, as a software program running on a personal computer or an application running on a smartphone, etc. The method according to the present disclosure may be provided to users in the form of SaaS (Software as a Service).

[1574] In the above embodiment, an example was given in which the specific processing is performed by one computer 22, but the technology of the present disclosure is not limited to this, and the specific processing may be distributed and performed by a plurality of computers including the computer 22. For example, the data generation model 58 may be provided in an external device of the data processing device 12, and data may be generated in the external device in accordance with input data.

[1575] In the above embodiment, an example in which the specific processing program 56 is stored in the storage 32 has been described, but the technology of the present disclosure is not limited to this. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-transitory storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-transitory storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes the specific processing in accordance with the specific processing program 56.

[1576] 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.

[1577] It is not necessary to store all of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store all of the specific processing program 56 in the storage 32; only a portion of the specific processing program 56 may be stored.

[1578] The hardware resource for executing a specific process can be any of the following processors: An example of a processor is a CPU, which is a general-purpose processor that functions as a hardware resource for executing a specific process by executing software, i.e., a program. Another example of a processor is a dedicated electrical circuit, such as an FPGA (Field-Programmable Gate Array), a PLD (Programmable Logic Device), or an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration designed specifically for executing a specific process. Each processor has built-in or connected memory, and each processor uses the memory to execute the specific process.

[1579] The hardware resource that executes the specific processing may be configured with one of these various processors, or may be configured with a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Also, the hardware resource that executes the specific processing may be a single processor.

[1580] As an example of a system configured with a single processor, first, one processor is configured by combining one or more CPUs and software, and this processor functions as a hardware resource that executes a specific process. Second, there is a system that uses a processor that realizes the functions of an entire system including multiple hardware resources that execute a specific process on a single IC chip, as typified by SoC (System-on-a-chip). In this way, a specific process is realized using one or more of the above-mentioned various processors as hardware resources.

[1581] Furthermore, the hardware structure of these various processors can be, more specifically, an electric circuit that combines circuit elements such as semiconductor devices. The specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps may be deleted, new steps may be added, or the processing order may be rearranged, without departing from the spirit of the invention.

[1582] The above-described description and illustrations are a detailed explanation of the parts related to the technology of the present disclosure and are merely an example of the technology of the present disclosure. For example, the above description of the configuration, functions, actions, and effects is an explanation of an example of the configuration, functions, actions, and effects of the parts related to the technology of the present disclosure. Therefore, it goes without saying that unnecessary parts may be deleted, new elements may be added, or replacements may be made to the above-described description and illustrations within the scope of the gist of the technology of the present disclosure. Furthermore, to avoid confusion and facilitate understanding of the parts related to the technology of the present disclosure, the above-described description and illustrations omit explanations of common technical knowledge that do not require particular explanation to enable the implementation of the technology of the present disclosure.

[1583] All publications, patent applications, and technical standards mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent application, or technical standard was specifically and individually indicated to be incorporated by reference.

[1584] The following is further disclosed regarding the above embodiment.

[1585] (Claim 1)

[1586] an input means for inputting the contents of an inquiry;

[1587] a refinement means for refining the query content received from the input means using a generative model;

[1588] a display means for displaying the inquiry content revised by the revision means;

[1589] A system including:

[1590] (Claim 2)

[1591] The system of claim 1, including embedding techniques to accommodate knowledge and terminology specific to the generative model.

[1592] (Claim 3)

[1593] 10. The system of claim 1, further comprising interactive means for enabling a user to select or modify the refinement suggestions generated by said refinement means.

[1594] "Example 1"

[1595] (Claim 1)

[1596] an input means for inputting the contents of an inquiry;

[1597] a refinement means for refining the query content received from the input means using a generative model;

[1598] a processing means for tokenizing query content for input to the generative model;

[1599] a return means for returning the inquiry content revised by the revision means in a JSON format;

[1600] a display means for displaying the inquiry content revised by the revision means;

[1601] A system including:

[1602] (Claim 2)

[1603] The system of claim 1, including embedding techniques to accommodate knowledge and terminology specific to the generative model.

[1604] (Claim 3)

[1605] 10. The system of claim 1, further comprising interactive means for enabling a user to select or modify the refinement suggestions generated by said refinement means.

[1606] "Application Example 1"

[1607] (Claim 1)

[1608] an input means for inputting the contents of an inquiry;

[1609] a refinement means for refining the query content received from the input means using a generative model;

[1610] a display means for displaying the inquiry content revised by the revision means;

[1611] A means to check the displayed revision proposal and send an inquiry.

[1612] A system including:

[1613] (Claim 2)

[1614] The system of claim 1, including embedding techniques to accommodate knowledge and terminology specific to the generative model.

[1615] (Claim 3)

[1616] 10. The system of claim 1, further comprising interactive means for enabling a user to select or modify the refinement suggestions generated by said refinement means.

[1617] "Example 2: Combining Emotion Engines"

[1618] (Claim 1)

[1619] an input means for inputting the contents of an inquiry;

[1620] a refinement means for refining the query content received from the input means using a generative model;

[1621] a display means for displaying the inquiry content revised by the revision means;

[1622] an emotion recognition means for analyzing emotions from the inquiry content generated by the refinement means;

[1623] an adjustment means for adjusting the content of the inquiry based on the emotion analyzed by the emotion recognition means;

[1624] A system including:

[1625] (Claim 2)

[1626] The system of claim 1, including embedding techniques to accommodate knowledge and terminology specific to the generative model.

[1627] (Claim 3)

[1628] 10. The system of claim 1, further comprising interactive means for enabling a user to review and modify the final query statement generated by said adjustment means.

[1629] "Application example 2 when combining emotion engines"

[1630] (Claim 1)

[1631] an input means for inputting the contents of an inquiry;

[1632] a refinement means for refining the query content received from the input means using a generative model;

[1633] a display means for displaying the inquiry content revised by the revision means;

[1634] an emotion recognition means for analyzing the emotion of the inquiry content;

[1635] an optimization means for optimizing query content based on the emotion analyzed by the emotion recognition means;

[1636] A system including:

[1637] (Claim 2)

[1638] The system of claim 1, including embedding techniques to accommodate knowledge and terminology specific to the generative model.

[1639] (Claim 3)

[1640] 10. The system of claim 1, further comprising interactive means for enabling a user to select or modify the refinement suggestions generated by said refinement means. [Explanation of symbols]

[1641] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Device 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robot< / url:> < / url:> < / url:> < / url:>

Claims

1. an input means for inputting the contents of an inquiry; a refinement means for refining the query content received from the input means using a generative model; a display means for displaying the inquiry content revised by the revision means; A system including:

2. The system of claim 1 , further comprising an embedding technique for accommodating knowledge and terminology specific to the generative model.

3. 2. The system of claim 1, further comprising interactive means for enabling a user to select or modify the refinement suggestions generated by said refinement means.

Images