Information processing apparatus, information processing system, information processing method, and program

Abstract

To reduce the workload of an administrator.MEANS FOR SOLVING THE PROBLEM: An information processing apparatus includes: an input reception unit for receiving an input of an instruction to generate registered question data and registered answer data; a request unit for requesting a large language model to generate registered answer data corresponding to the registered question data on the basis of the generation instruction; and an output unit for outputting, to a terminal device, the registered answer data output from the large language model in response to the request.SELECTED DRAWING: Figure 9

Description

[Technical Field]

[0001] The present invention relates to an information processing device, an information processing system, an information processing method, and a program. [Background technology]

[0002] In recent years, chatbot systems have become popular, which provide answers to questions entered by users. It is known that such chatbot systems prepare question and answer information that associates questions with answers, and provide answers to entered questions based on the question and answer information. Summary of the Invention [Problem to be solved by the invention]

[0003] In the conventional chatbot systems described above, the administrator of the chatbot system must manually input a large number of questions and answers to provide valid answers to questions. This input work takes a lot of time, especially when introducing a new chatbot system, and places a heavy burden on the administrator.

[0004] The disclosed technology has been developed in light of the above circumstances, and aims to reduce the burden on administrators of manually inputting a large number of questions and answers. [Means for solving the problem]

[0005] The disclosed technology is an information processing device having an input receiving unit that receives input of registered question data and instructions to generate registered answer data, a request unit that outputs to a large-scale language model the generation of registered answer data corresponding to the registered question data based on the generation instructions, and an output unit that outputs the registered answer data output from the large-scale language model in response to the request to a terminal device. [Effects of the Invention]

[0006] This reduces the burden on administrators of manually entering a large number of questions and answers. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 illustrates an example of a system configuration of a chat system. [Figure 2] FIG. 2 illustrates an example of a hardware configuration of an information processing device. [Figure 3] FIG. 2 illustrates an example of a hardware configuration of a terminal device. [Figure 4] FIG. 2 is a diagram illustrating an example of a conversation history storage unit. [Figure 5] FIG. 2 illustrates an example of a question and answer storage unit. [Figure 6] FIG. 2 illustrates an example of a document storage unit. [Figure 7] FIG. 2 illustrates an example of a user information storage unit. [Figure 8] FIG. 2 is a diagram illustrating the functions of each device in the chat system of the first embodiment. [Figure 9] FIG. 2 is a first sequence diagram illustrating the operation of the chat system according to the first embodiment. [Figure 10] FIG. 10 is a second sequence diagram illustrating the operation of the chat system according to the first embodiment. [Figure 11] FIG. 10 is a first diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 12] FIG. 10 is a first diagram illustrating a prompt according to the first embodiment. [Figure 13] FIG. 10 is a second diagram illustrating the prompt according to the first embodiment. [Figure 14] FIG. 10 is a second diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 15] FIG. 10 is a third diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 16] FIG. 10 is a fourth diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 17] FIG. 10 is a fifth diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 18]FIG. 6 is a sixth diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 19] FIG. 11 is a first diagram illustrating a display example of a terminal device according to the second embodiment. [Figure 20] FIG. 10 is a first diagram illustrating a prompt according to the second embodiment. [Figure 21] FIG. 11 is a second diagram illustrating a display example of the terminal device according to the second embodiment. [Figure 22] FIG. 10 is a second diagram illustrating the prompt according to the second embodiment. [Figure 23] FIG. 11 is a first diagram illustrating a display example of a terminal device according to the third embodiment. [Figure 24] FIG. 10 is a diagram illustrating a prompt according to the third embodiment. [Figure 25] FIG. 10 is a diagram illustrating another prompt according to the third embodiment. [Figure 26] FIG. 11 is a second diagram illustrating a display example of the terminal device according to the third embodiment. [Figure 27] FIG. 10 illustrates an example of a system configuration of a chat system according to a fourth embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] (First embodiment) The present embodiment will be described below with reference to the drawings: Fig. 1 is a diagram showing an example of the system configuration of a chat system.

[0009] The chat system 100 of this embodiment is an example of an information processing system that includes an information processing device 200, a first storage device 300, a second storage device 400, a server device 500, a terminal device 600, and a terminal device 700, each of which is connected via a communication network 1.

[0010] In the chat system 100, the information processing device 200, the first storage device 300, the second storage device 400, and the server device 500 provide a chatbot service to the user of the terminal device 600.

[0011] When a question is input from the terminal device 600, the information processing device 200 references the question and answer information stored in the first storage device 300 and outputs an answer to the question to the terminal device 600.

[0012] The first storage device 300 has a conversation history storage unit 310 and a question and answer storage unit 320. The conversation history storage unit 310 stores conversation history information that associates questions input from the terminal device 600 with answers output to the terminal device 600. The conversation history information may be managed for each tenant that introduces the chat system 100. The tenant in this embodiment may be an organization such as a company, or may be the organization to which a user who uses the chat system 100 belongs.

[0013] The question and answer storage unit 320 stores question and answer information in which questions and answers are associated with each other. In other words, the question and answer storage unit 320 is an example of a question and answer information group in which question data indicating a question is associated with answer data indicating an answer. The question and answer information is information in which question data indicating a question is associated with answer data indicating an answer in pairs. The question and answer information is information that is stored in advance in the question and answer storage unit 320, and may be stored for each tenant that introduces the chat system 100.

[0014] In the following description, question data included in the question and answer information stored in the question and answer storage unit 320 will be referred to as registered question data, and answer data included in the question and answer information will be referred to as registered answer data.

[0015] The registered question data and registered answer data in this embodiment may be generated by a large-scale language model provided in the server device 500. Furthermore, the registered question data and registered answer data generated by the large-scale language model provided in the server device 500 in this embodiment may be checked and edited by an administrator of the chat system 100.

[0016] The second storage device 400 has a document storage unit 410 and a user information storage unit 420. The document storage unit 410 stores user-specific information that is not publicly available.

[0017] A user of chat system 100 may be, for example, a tenant, in which case the user-specific information is tenant-specific information. A user of chat system 100 may also be an individual user, in which case the user-specific information is user-specific information.

[0018] In the following description, user-specific information will be referred to as in-house information, and in-house information will be taken as an example of user-specific information.

[0019] In the present embodiment, internal company information may be stored for each tenant in the document storage unit 410. The user information storage unit 420 stores, for each tenant, user information relating to users who belong to the tenant.

[0020] The server device 500 has a large-scale language model storage unit 510 that stores a large-scale language model, and realizes a generative AI (Artificial Intelligence). The generative AI generates and outputs text data in response to an input prompt, such as ChatGPT. In the following description, data output from the large-scale language model may be referred to as output data.

[0021] The server device 500 of this embodiment transmits output data output by the large-scale language model.

[0022] Furthermore, the prompts of this embodiment are text data for expressing instructions or questions to the generation AI realized by the server device 500. Details of the large-scale language model will be described later.

[0023] The terminal device 600 is mainly used by a user who uses the chat system 100. More specifically, the terminal device 600 inputs question data indicating a question to the chat system 100 to the information processing device 200. The terminal device 600 also displays answer data output from the information processing device 200, which indicates an answer to the question indicated by the input question data, in association with the input question data.

[0024] In the following description, question data transmitted from the terminal device 600 to the information processing device 200 and input to the information processing device 200 will be referred to as input question data. Also, in the following description, answer data output from the information processing device 200 to the terminal device 600 will be referred to as output answer data.

[0025] The terminal device 700 is mainly used by an administrator who manages the chat system 100. More specifically, the terminal device 700 performs operations such as generating registered question data and registered answer data to be stored in the question and answer storage unit 320.

[0026] In this embodiment, each of the terminal devices 600 and 700 may be, for example, a smartphone or a tablet terminal.

[0027] In the chat system 100 of this embodiment, when registered question data and information to be used for generating the registered answer data are specified, the information processing device 200 outputs a prompt to the server device 500, requesting the server device 500 to generate the registered question data and the registered answer data corresponding to the registered question data, from the specified information. In response to the prompt, the information processing device 200 associates the registered question data and the registered answer data output from the large-scale language model to generate question answer information, and causes the terminal device 700 to display a screen for confirming and editing the content of the question answer information. The information processing device 200 then stores the question answer information, after the content of the question answer information has been confirmed and edited on the terminal device 700, in the question and answer storage unit 320.

[0028] In this embodiment, the registered question data and registered answer data are generated in the large-scale language model 511 included in the server device 500. Therefore, according to this embodiment, the administrator of the chat system 100 does not need to manually input the registered question data and registered answer data, thereby reducing the burden on the administrator.

[0029] Furthermore, in this embodiment, a screen that allows the administrator of the chat system 100 to confirm the content is displayed on the terminal device 700. Therefore, according to this embodiment, the administrator can confirm and edit the content of the question and answer information generated by the large-scale language model. Furthermore, according to this embodiment, question and answer information with appropriate content can be stored in the question and answer storage unit 320, and output of output answer data with incorrect content in response to input question data entered by a user can be prevented.

[0030] 1, the information processing device 200, the first storage device 300, and the second storage device 400 are each separate devices, but this is not limiting. The first storage device 300 and the second storage device 400 may be included in the information processing device 200.

[0031] In addition, in FIG. 1, the information processing device 200 may be realized by one or more information processing devices, and may be realized by a plurality of information processing devices.

[0032] Furthermore, the chat system 100 only needs to include the information processing device 200 and the terminal device 700, and other devices may not be included in the chat system 100.

[0033] Next, an example of the hardware configuration of each device included in the chat system 100 of this embodiment will be described.

[0034] In this embodiment, the information processing device 200, the first storage device 300, the second storage device 400, and the server device 500 are general-purpose computers, and therefore, in FIG. 2, the hardware configuration of the information processing device 200 will be described as an example.

[0035] In this embodiment, the terminal devices 600 and 700 are smartphones, and in FIG. 3, the hardware configuration of the terminal device 700 will be described as an example.

[0036] Fig. 2 is a diagram showing an example of the hardware configuration of an information processing device. The information processing device 200 shown in Fig. 2 is constructed by a computer (information processing device), and includes a CPU 201, a ROM 202, a RAM 203, an HD 204, an HDD (Hard Disk Drive) controller 205, a display 206, an external device connection I / F (Interface) 208, a network I / F 209, a bus line 210, a keyboard 211, a pointing device 212, a DVD-RW (Digital Versatile Disk Rewritable) drive 214, and a media I / F 216, as shown in Fig. 2.

[0037] Of these, the CPU 201 controls the overall operation of the information processing device 200. The ROM 202 stores programs used to drive the CPU 201, such as an IPL. The RAM 203 is used as a work area for the CPU 501. The HD 204 stores various data, such as programs. The HDD controller 205 controls the reading and writing of various data from and to the HD 204 under the control of the CPU 501. The display 206 displays various information, such as a cursor, menus, windows, characters, or images. The external device connection I / F 208 is an interface for connecting various external devices. In this case, the external devices are, for example, USB (Universal Serial Bus) memories, printers, etc. The network I / F 209 is an interface for data communication using a communication network. The bus line 210 is an address bus, a data bus, etc. for electrically connecting the components, such as the CPU 201, shown in FIG. 2.

[0038] The keyboard 211 is a type of input means having multiple keys for inputting characters, numbers, various instructions, etc. The pointing device 212 is a type of input means for selecting and executing various instructions, selecting a processing target, moving a cursor, etc. The DVD-RW drive 214 controls reading and writing of various data from and to a DVD-RW 213, which is an example of a removable recording medium. Note that this is not limited to a DVD-RW, and may be a DVD-R, etc. The media I / F 216 controls reading and writing (storing) of data from and to a recording medium 215, such as a flash memory.

[0039] 3 is a diagram showing an example of the hardware configuration of a terminal device 700 according to this embodiment. The terminal device 700 includes a CPU 701, a ROM 702, a RAM 703, an EEPROM 704, a CMOS sensor 705, an image sensor I / F 706, an acceleration / direction sensor 707, a media I / F 709, and a GPS receiver 711.

[0040] Of these, the CPU 701 is an arithmetic processing unit that controls the overall operation of the terminal device 700. The ROM 702 stores the CPU 701 and programs used to drive the CPU 701, such as the IPL. The RAM 703 is used as a work area for the CPU 701. The EEPROM 704 reads or writes various data, such as smartphone programs, under the control of the CPU 701. The ROM 702, RAM 703, and EEPROM 704 are examples of storage devices of the terminal device 700.

[0041] The CMOS (Complementary Metal Oxide Semiconductor) sensor 705 is a type of built-in imaging means that captures an image of a subject (mainly a self-portrait) and obtains image data under the control of the CPU 701. Note that instead of a CMOS sensor, an imaging means such as a CCD (Charge Coupled Device) sensor may also be used.

[0042] The imaging element I / F 706 is a circuit that controls the driving of the CMOS sensor 705. The acceleration / azimuth sensor 707 is a variety of sensors, such as an electronic magnetic compass that detects geomagnetism, a gyrocompass, and an acceleration sensor. The media I / F 709 controls the reading and writing (storage) of data from and to a recording medium 708, such as a flash memory. The GPS receiver 711 receives GPS signals from GPS satellites.

[0043] The terminal device 700 also includes a long-distance communication circuit 712, an antenna 712a of the long-distance communication circuit 712, a CMOS sensor 713, an image sensor I / F 714, a microphone 715, a speaker 716, an audio input / output I / F 717, a display 718, an external device connection I / F (Interface) 719, a short-distance communication circuit 720, an antenna 720a of the short-distance communication circuit 720, and a touch panel 721.

[0044] Of these, the long-distance communication circuit 712 is a circuit that communicates with other devices via a communication network. The CMOS sensor 713 is a type of built-in imaging means that captures an image of a subject and obtains image data under the control of the CPU 701. The image sensor I / F 714 is a circuit that controls the driving of the CMOS sensor 713. The microphone 715 is a built-in circuit that converts sound into an electrical signal. The speaker 716 is a built-in circuit that converts an electrical signal into physical vibrations to produce sound such as music or voice. The sound input / output I / F 717 is a circuit that processes the input and output of sound signals between the microphone 715 and the speaker 716 under the control of the CPU 701.

[0045] The display 718 is a type of display means such as a liquid crystal display or organic electroluminescence (EL) display that displays an image of a subject, various icons, etc. The external device connection I / F 719 is an interface for connecting various external devices. The short-range communication circuit 720 is a communication circuit such as NFC (Near Field Communication) or Bluetooth (registered trademark). The touch panel 721 is a type of input means that allows a user to operate the terminal device 700 by pressing the display 718. The display 718 is an example of a display unit that the terminal device 700 has.

[0046] 3 is an example of the terminal device 700, and the hardware configuration of the terminal device 700 is not limited to this. The hardware configuration of the terminal device 700 may be the configuration shown in FIG. 2. The hardware configuration of the terminal device 600 may be the same as that of the terminal device 700, and therefore a description thereof will be omitted.

[0047] Next, each storage unit will be described with reference to Figures 4 to 7. Figure 4 is a diagram showing an example of a conversation history storage unit.

[0048] The conversation history information stored in the conversation history storage unit 310 shown in FIG. 4 includes items such as a conversation ID, date and time, a tenant ID, a question, and an answer, and is information in which values ​​of each item are associated with each other.

[0049] The value of the item "Conversation ID" is identification information for identifying a conversation. The value of the item "Date and Time" indicates the date and time when input question data was input to the terminal device 600 or the date and time when output answer data was output to the terminal device 600. The value of the item "Tenant ID" is identification information for identifying the tenant to which the user of the terminal device 600 belongs.

[0050] The value of the item "Question" indicates input question data that is input as a question sentence to the terminal device 600. The value of the item "Answer" indicates output answer data that is output to the terminal device 600 as an answer sentence.

[0051] In this way, the conversation history information in this embodiment includes a conversation ID that identifies the conversation history information, the tenant ID of the tenant to which the user of the terminal device 600 belongs, the date and time when the input question data was input or the date and time when the output answer data was output, the input question data, and the output answer data.

[0052] The conversation history information shown in FIG. 4 is an example, and the conversation history information of this embodiment may include items other than the items of information shown in FIG.

[0053] Fig. 5 is a diagram illustrating an example of a question and answer storage unit. The question and answer information stored in the question and answer storage unit 320 of this embodiment may be provided for each tenant and managed by each tenant. The example of Fig. 5 shows question and answer information managed by a tenant identified by tenant ID "11."

[0054] The question and answer information includes information items such as tenant ID, question and answer ID, update date and time, question, answer, contact information, and link, and the items "tenant ID" and "question and answer ID" are associated with other items. The question and answer information of this embodiment is information in which the value of the item "question and answer ID" is associated with the values ​​of the other items.

[0055] The value of the item "Question and Answer ID" indicates identification information that identifies the question and answer information. The value of the item "Update Date and Time" indicates the date and time when the question and answer information was updated. The value of the item "Question" is the registered question data (text data), and the value of the item "Answer" is the registered answer data (text data). The values ​​of the items "Question" and "Answer" are stored in advance by the administrator of each tenant, etc.

[0056] The value of the item "Contact Information" indicates the contact information to be used when the user of the terminal device 600 makes an inquiry related to the registered question data. The contact information may be, for example, an email address or a telephone number, and may be stored in the user information storage unit 420.

[0057] The value of the item "Link" indicates a link that shows information related to the registered question data or registered answer data.

[0058] The items "contact information" and "links" do not have to be stored in the question and answer storage unit 320, as long as the question and answer storage unit 320 stores registered question data and registered answer data.

[0059] Fig. 6 is a diagram illustrating an example of a document storage unit. In the document storage unit 410 of this embodiment, internal company information may be stored for each tenant and managed for each tenant. Fig. 6 illustrates an example of internal company information managed by a tenant identified by tenant ID "11."

[0060] In the document storage unit 410 of this embodiment, the internal company information is divided into predetermined units and each divided piece of information is stored. In the following description, each divided piece of internal company information is called a chunk. In the document storage unit 410, the chunk may be stored in association with identification information that identifies the chunk.

[0061] 6, one page of internal information in an internal document is defined as a chunk, and the page number is used as identification information to identify the chunk. Note that a chunk is not limited to one page of internal information, and may be, for example, each paragraph or each chapter in an internal document.

[0062] 7 is a diagram illustrating an example of the user information storage unit 420. The user information stored in the user information storage unit 420 of this embodiment is in-house information stored for each tenant, and may be managed for each tenant.

[0063] The user information of this embodiment includes information items such as a tenant ID, a user ID, a user name, and an email address. The value of the item "tenant ID" is identification information that identifies the tenant to which the user identified by the user ID belongs. The value of the item "user ID" is identification information that identifies the user of the terminal device 600. The item "user name" indicates the name of the user identified by the user ID, and the value of the item "email address" indicates the email address of the user identified by the user ID.

[0064] In this embodiment, the user information storage unit 420 is stored in the second storage device 400, but the user information storage unit 420 does not have to be provided in the second storage device 400, and it is sufficient if only the document storage unit 410 is provided.

[0065] Next, the functional configuration of each device included in the chat system 100 of this embodiment will be described with reference to Fig. 8. Fig. 8 is a diagram illustrating the functions of each device included in the chat system of the first embodiment.

[0066] First, a description will be given of the functional configuration of the information processing device 200. The information processing device 200 of this embodiment includes an input receiving unit 251, a storage control unit 252, a setting reference unit 253, an information acquisition unit 254, a prompt generation unit 255, a request unit 256, a list generation unit 257, an output unit 258, an editing unit 259, and a question and answer unit 260.

[0067] The input receiving unit 251 receives various inputs to the information processing device 200. Specifically, the input receiving unit 251 receives input of various information input from the terminal device 700 and input of input question data from the terminal device 600. Note that the various information input from the terminal device 700 includes, for example, in-house information stored in the document storage unit 410, information identifying the in-house information used to generate the registered question data and the registered answer data, setting information set when the registered question data and the registered answer data are generated, and the like.

[0068] The storage control unit 252 controls how the in-house information is stored in the document storage unit 410 of the second storage device 400. Specifically, the storage control unit 252 divides the in-house information received from the terminal device 700 into chunks, and stores the in-house document for each chunk in the document storage unit 410.

[0069] Specifically, for example, if the internal company information is document information consisting of multiple pages, the storage control unit 252 divides the internal company information into chunks, with one page as a chunk, and stores the chunks in the document storage unit 410 by associating them with identification information that identifies the chunks.

[0070] The setting reference unit 253 references setting information related to the generation of registered question data and registered answer data from the information received by the input receiving unit 251 from the terminal device 700 .

[0071] The information acquisition unit 254 acquires the in-house information from the document storage unit 410. Specifically, when the input reception unit 251 receives an input of an instruction to generate registered question data and registered answer data, the information acquisition unit 254 acquires the in-house information stored in the document storage unit 410 of the second storage device 400 for each chunk.

[0072] The prompt generation unit 255 generates a prompt to be output to the server device 500. For example, when the information acquisition unit 254 acquires in-house information chunk by chunk, the prompt generation unit 255 generates registration question data using the acquired in-house information and a prompt requesting the generation of registration answer data. Furthermore, when the setting reference unit 253 references setting information, the prompt generation unit 255 generates registration question data according to the setting indicated by the setting information and a prompt requesting the generation of registration answer data.

[0073] Furthermore, the prompt generated by the prompt generating unit 255 requests the server device 500 to output the generated registration question data and registration answer data as data in JSON format.

[0074] The request unit 256 outputs the prompt generated by the prompt generation unit 255 to the server device 500. In other words, the request unit 256 makes a request based on the prompt to the large-scale language model 511. In this embodiment, making a request based on the prompt to the large-scale language model 511 may be expressed as sending the prompt generated by the prompt generation unit 255 to the server device 500.

[0075] The list generation unit 257 generates a list of pairs of registered question data and registered answer data based on the output data output from the server device 500 to the information processing device 200. A pair of registered question data and registered answer data indicates registered question data and registered answer data associated with this registered question data.

[0076] The output unit 258 outputs various types of information from the information processing device 200. Specifically, for example, the output unit 258 outputs information indicating a list of pairs of registered question data and registered answer data generated by the list generation unit 257 to the terminal device 700. The output unit 258 also outputs the registered question data and registered answer data stored in the question and answer storage unit 320 to the first storage device 300. The output unit 258 also outputs, for example, internal company information stored in the document storage unit 410.

[0077] The editing unit 259 edits the question and answer information stored in the question and answer storage unit 320. More specifically, when the editing unit 259 receives an editing operation on the registered question data or registered answer data at the terminal device 700, the editing unit 259 reflects the editing operation in the question and answer information stored in the question and answer storage unit 320. In other words, the editing unit 259 overwrites the registered question data or registered answer data stored in the question and answer storage unit 320 in accordance with the editing operation by the administrator of the chat system 100.

[0078] The question and answering unit 260 provides a chatbot system using the question and answer information stored in the question and answer storage unit 320.

[0079] Specifically, when input question data is input from the terminal device 600, the question answering unit 260 outputs a search request using the input question data to the first storage device 300, and causes the search unit 330 of the first storage device 300 to search the question and answer storage unit 320 for the input question data. The question answering unit 260 also acquires registered answer data that is the search result of the question and answer storage unit 320 from the first storage device 300. The question answering unit 260 also causes the output unit 258 to output the registered answer data acquired from the first storage device 300 to the terminal device 600 as output answer data.

[0080] Next, we will explain the functions of the first storage device 300. The first storage device 300 includes a conversation history storage unit 310, a question and answer storage unit 320, a search unit 330, and a storage unit 340. The search unit 330 searches the question and answer storage unit 320 in response to a search request from the information processing device 200, and transmits the search results to the information processing device 200.

[0081] The storage unit 340 stores the registered question data and registered answer data output from the information processing device 200 in the question and answer storage unit 320. Furthermore, when input question data is input and output answer data is output in the information processing device 200, the storage unit 340 stores conversation history information including the input question data and the output answer data in the conversation history storage unit 310.

[0082] Next, the functions of the second storage device 400 will be described. The second storage device 400 includes a document storage unit 410, a user information storage unit 420, a search unit 430, and a storage unit 440. The search unit 430 searches the document storage unit 410 and the user information storage unit 420 in response to a search request from the information processing device 200, and transmits the search results to the information processing device 200. When an administrator of the chat system 100 or the like performs an operation to upload in-house information, the storage unit 440 stores the new in-house information in the document storage unit 410.

[0083] Next, we will explain the functions of the server device 500. The server device 500 has a large-scale language model storage unit 510, an input unit 520, and an output unit 530. The large-scale language model storage unit 510 stores a large-scale language model 511 that realizes a generative AI.

[0084] The large-scale language model 511 of this embodiment is a computer language model that is generated by executing a learning process using a huge amount of unlabeled text as learning data and is composed of an artificial neural network with a large number of parameters. The large-scale language model 511 can capture much of the syntax and meaning of human language by being sufficiently trained using a context learning technique such as next sentence prediction, which understands the context by determining whether sentence 1 and sentence 2 are consecutive, or a masked language model, which understands the context by masking words in a sentence and predicting the masked words from the words before and after them.

[0085] The input unit 520 inputs a prompt transmitted from the information processing device 200 to the large-scale language model 511. The output unit 530 outputs (transmits) output data output from the large-scale language model 511 to the information processing device 200. In other words, the output unit 530 transmits output response data output from the large-scale language model 511 to the information processing device 200.

[0086] Next, a description will be given of the functions of the terminal devices 600 and 700. The terminal devices 600 and 700 each include a communication control unit 630 and 730, an input receiving unit 640 and 740, and a display control unit 650 and 750.

[0087] The communication control units 630, 730 control communication between the terminal devices 600, 700, respectively, and the information processing device 200. The input receiving units 640, 740 receive input to the terminal devices 600, 700. The display control units 650, 750 control display on the terminal devices 600, 700.

[0088] Next, the operation of the chat system 100 of this embodiment will be described with reference to Fig. 9. Fig. 9 is a first sequence diagram illustrating the operation of the chat system of the first embodiment.

[0089] 9 and 10 are operations performed when the administrator of the chat system 100 has logged in to the chat system 100 using his / her own user ID and the tenant ID to which he / she belongs, etc. In other words, the operations shown in Fig. 9 and 10 are operations performed when the administrator of the chat system 100 is able to access the document storage unit 410 and the question and answer storage unit 320 associated with the tenant ID to which he / she belongs.

[0090] 9, steps S901 to S906 show the operation of storing in-house information specified on the terminal device 700 in the document storage unit 410 of the second storage device 400. Also, steps S907 to S920 show the operation of generating registered question data and registered answer data and displaying the generated results on the terminal device 700.

[0091] In this embodiment, the operations from step S901 to step S906 and the operations from step S907 to step S920 may be performed at different, independent times.

[0092] In the chat system 100, the terminal device 700 displays a generation screen for generating registered question data and registered answer data (step S901). The generation screen may be displayed on the display 718 of the terminal device 700 in response to an operation on the terminal device 700 by an administrator of the chat system 100.

[0093] Next, the terminal device 700 receives input of information and setting information used to generate the registered question data and registered answer data via the input receiving unit 740 (step S902). The information input here includes in-house information stored in the document storage unit 410 and setting information related to the generation of the registered question data and registered answer data.

[0094] Next, the terminal device 700 transmits the input information to the information processing device 200 via the communication control unit 730 (step S903). The information transmitted to the information processing device 200 here includes the in-house information stored in the document storage unit 410 and setting information related to the generation of registered question data and registered answer data.

[0095] When the information processing device 200 acquires information from the terminal device 700, the storage control unit 252 divides the in-house information contained in the information acquired from the terminal device 700 into chunks and associates each chunk with identification information that identifies the chunk (step S904).

[0096] Next, the information processing device 200 outputs the in-house information divided into chunks to the second storage device 400 via the output unit 258 (step S905).

[0097] When the second storage device 400 receives the internal company information for each chunk from the information processing device 200, the storage unit 440 stores the internal company information for each chunk in the document storage unit 410 associated with the tenant ID of the tenant to which the administrator of the chat system 100 belongs (step S906).

[0098] Furthermore, when the terminal device 700 receives an operation to instruct the generation of registered question data and registered answer data via the input receiving unit 740, the terminal device 700 transmits an instruction to generate the registered question data and registered answer data via the communication control unit 730 to the information processing device 200 (step S907). Note that this instruction to generate may include information that identifies in-house information to be used when generating the registered question data and registered answer data. In other words, the instruction to generate may include information that specifies in-house information that serves as the source of the registered question data and registered answer data.

[0099] When the information processing device 200 receives this generation instruction via the input receiving unit 251, the information acquiring unit 254 sends a request to acquire chunks included in the internal company information used to generate the registered question data and the registered answer data, from the internal company information identified by the information identifying the internal company information included in the generation instruction (step S908). Note that this acquisition request may include information identifying the internal company information used to generate the registered question data and the registered answer data.

[0100] The second storage device 400 receives the acquisition request, identifies the in-house information to be sent to the information processing device 200 based on the information identifying the in-house information included in the acquisition request, and acquires the chunks included in the identified in-house information (step S909). Subsequently, the second storage device 400 transmits the acquired chunks to the information processing device 200 (step S910).

[0101] When the information processing device 200 acquires the chunk, the prompt generation unit 255 generates a pair of registered question data and registered answer data using the acquired chunk, and generates a prompt requesting that the data be output as JSON format data (step S911).

[0102] JSON format data is data in a format that allows the information processing device 200 to automatically read and link registered question data and registered answer data.

[0103] In other words, the prompt generation unit 255 of this embodiment generates a prompt that requests that the results of generating a set of registered question data and registered answer data using the specified information be output in a format that can be read by the information processing device 200. The prompt generated by the prompt generation unit 255 will be described in detail later.

[0104] Next, the information processing device 200 causes the request unit 256 to output the prompt generated in step S911 to the server device 500 (step S912). In other words, the information processing device 200 causes the request unit 256 to make a request indicated by the prompt generated in step S911 to the large-scale language model 511. The server device 500 causes the input unit 520 to input the prompt to the large-scale language model 511, and the output unit 530 to transmit output data output from the large-scale language model 511 to the information processing device 200 (step S913).

[0105] The output data output from the server device 500 to the information processing device 200 in step S913 is JSON format data that includes a set of registered question data and registered answer data using the chunks acquired in step S909 by the information processing device 200. Details of the output data output from the server device 500 will be described later.

[0106] In the chat system 100 of this embodiment, the operations from step S908 to step S913 are repeated until all chunks included in the in-house information identified in step S907 are acquired.

[0107] Next, the setting reference unit 253 of the information processing device 200 refers to the setting information input from the terminal device 700 in step S903 (step S914).

[0108] If the number of pairs of registered question data and registered answer data has not been set in step S914, information processing device 200 proceeds to step S918, which will be described later.

[0109] If the number of pairs of registered question data and registered answer data is set in step S914, the information processing device 200 causes the prompt generation unit 255 to generate a prompt requesting that the number of pairs of registered question data and registered answer data be set to the set number (step S915). Subsequently, the information processing device 200 causes the request unit 256 to output the prompt generated in step S915 to the server device 500 (step S916).

[0110] The server device 500 inputs the prompt to the large-scale language model 511, and transmits the output data output from the large-scale language model 511 to the information processing device 200 by the output unit 530. (step S917).

[0111] The output data output from the server device 500 to the information processing device 200 in step S917 is data in JSON format that includes the number of pairs of registered question data and registered answer data set according to the setting information.

[0112] Next, the information processing device 200 causes the list generating unit 257 to generate a list of pairs of registered question data and registered answer data included in the output data acquired from the server device 500 in step S913 or step S917 (step S918).

[0113] Specifically, the list generation unit 257 reads the registered question data and registered answer data from the output data generated in JSON format, and stores the registered question data and registered answer data in variables, thereby generating information indicating a list of pairs of registered question data and registered answer data.

[0114] In step S918, the set of registered question data and registered answer data may be temporarily stored in question and answer storage unit 320.

[0115] Next, the information processing device 200 causes the output unit 258 to output information indicating the list generated in step S918 to the terminal device 700 (step S919).

[0116] The terminal device 700 displays on the display 718 a list screen including information indicating the list acquired from the information processing device 200 in step S919 (step S920).

[0117] Next, with reference to FIG. 10, the operation of the chat system 100 after the list of pairs of registered question data and registered answer data is displayed on the terminal device 700 will be described.

[0118] FIG. 14 is a second flowchart illustrating the operation of the chat system of the first embodiment.

[0119] In the terminal device 700, the input receiving unit 740 receives an operation to select a pair of registered question data and registered answer data whose contents are to be confirmed on the list screen displayed on the display 718 (step S1001). Next, the terminal device 700 transmits information indicating the selected pair to the information processing device 200 via the communication control unit 730 (step S1002).

[0120] The information processing device 200, through the editing unit 259, acquires the registered question data and registered answer data included in the selected pair from the pairs of registered question data and registered answer data temporarily stored in the question and answer storage unit 320 (step S1003). Subsequently, the information processing device 200, through the editing unit 259, transmits an instruction to the terminal device 700 to display an editing screen for the acquired registered question data and registered answer data (step S1004).

[0121] The terminal device 700 causes the display control unit 750 to display an editing screen for the pair of registered question data and registered answer data selected in step S1001 (step S1005). Next, if an editing operation is input via the input receiving unit 740, the terminal device 700 performs editing in accordance with this operation (step S1006).

[0122] Next, the terminal device 700 receives an instruction to register the pair of question data and answer data (step S1007) and transmits a registration request to the information processing device 200 (step S1008). Note that this registration request includes the pair of question data and answer data after the administrator of the chat system 100 has confirmed the contents of the pair.

[0123] The registered question data and registered answer data after the administrator of the chat system 100 has confirmed the content include edited registered question data and registered answer data after editing operations have been performed on the editing screen, and registered question data and registered answer data for which no editing has been performed and only the content has been confirmed.

[0124] When the information processing device 200 receives a registration request through the input receiving unit 251, the editing unit 259 outputs a set of registration question data and registration answer data after the content has been confirmed to the first storage device 300 (step S1009).

[0125] When the first storage device 300 acquires a pair of registered question data and registered answer data from the information processing device 200, it assigns a question answer ID to the pair and stores them in the question and answer storage unit 320 as question answer information that associates the pair of registered question data and registered answer data with the update dates and times of the registered question data and registered answer data (step S1010).

[0126] The process of assigning the question and answer ID and associating the question and answer ID with the update date and time with the set of registered question data and registered answer data to create question and answer information may be performed by the information processing device 200.

[0127] Hereinafter, examples of prompts generated by the information processing device 200 and displays on the terminal device 700 will be described with reference to FIGS.

[0128] 11 is a first diagram showing a display example of the terminal device of the first embodiment. A screen 110 shown in Fig. 11 is an example of a generated screen displayed on the terminal device 700 in step S901 of Fig. 9.

[0129] The screen 110 includes an input area 111, a display area 112, and operation buttons 113. The input area 111 includes input fields 114 and 115 for inputting settings related to the generation of registered question data and registered answer data.

[0130] Input fields 114 and 115 are input fields for inputting the number of pairs of registered question data and registered answer data. Specifically, input field 114 is an input field for inputting the number of pairs of registered question data and registered answer data, and input field 115 is an input field for inputting that there is no limit on the number of pairs of registered answer data and registered answer data.

[0131] Display area 112 displays information 112a that identifies in-house information to be stored in document storage unit 410. Display area 112 also includes operation buttons .

[0132] In this embodiment, when the operation button 116 is operated after the information 112a specifying the in-house information is displayed in the display area 112, the in-house information specified by the information 112a specifying the in-house information is transmitted (uploaded) to the information processing device 200. Note that the information 112a specifying the in-house information may be, for example, the file name of the in-house information.

[0133] Operation button 113 is an operation button for instructing the generation of a set of registered question data and registered answer data.

[0134] When the operation button 113 is operated after the setting information is input and the company information is uploaded, the terminal device 700 sends an instruction to the information processing device 200 to generate a set of registered question data and registered answer data using the uploaded company information in accordance with the setting information.

[0135] In this embodiment, the information processing device 200 displays both the setting information and information identifying the internal company information used to generate the registered question data and the registered answer data, thereby enabling the number of pairs of registered question data and registered answer data to be generated to be set for each piece of internal company information.

[0136] Furthermore, in this embodiment, when storing in-house information in the document storage unit 410 , a set of registered question data and registered answer data can be generated using this in-house information and stored in the question and answer storage unit 320 .

[0137] Furthermore, according to this embodiment, the administrator of the chat system 100 can simply input setting information and select the internal company information to upload, and generate a set of registered question data and registered answer data using this internal company information as the information source, and store the data in the question and answer storage unit 320.

[0138] Therefore, according to this embodiment, the burden on the administrator of the chat system 100 can be reduced.

[0139] In the example of FIG. 11, the information referenced when generating a set of registered question data and registered answer data is the in-house information identified by the information 112a that identifies the in-house information, but this is not limiting.

[0140] The information referenced when generating a pair of registered question data and registered answer data is not limited to, and does not have to be, internal company information. In other words, the information referenced when generating a pair of registered question data and registered answer data may be information other than user-specific information.

[0141] 11, the information 112a identifying the in-house information is the file name of the in-house information, but is not limited to this. The information 112a identifying the in-house information may be, for example, a link indicating the location where the in-house information is stored.

[0142] In this embodiment, when the operation button 113 on the screen 110 is operated, the information processing device 200 retrieves the internal company information stored in the document storage unit 410 chunk by chunk, and generates a prompt requesting the generation of a pair of registered question data and registered answer data by the prompt generation unit 255.

[0143] The prompt for requesting the generation of a set of registered question data and registered answer data will be described below with reference to Fig. 12. Fig. 12 is a first diagram illustrating the prompt of the first embodiment.

[0144] The prompt 12 shown in Fig. 12 is generated by the prompt generation unit 255 in step S911 of Fig. 9. In other words, the prompt 12 shown in Fig. 12 is a prompt requesting that a set of registered question data and registered answer data using chunks included in the company information identified on screen 110 be output as JSON format data.

[0145] The prompt 12 includes text data 12a, 12b, and 12c. The text data 12a is text data that indicates to the large-scale language model 511 that the role of the large-scale language model 511 is to generate registered question data and registered answer data, and that it will generate pairs of registered question data and registered answer data and output them as data in JSON format.

[0146] The text data 12b is an information source used to generate a pair of registered question data and registered answer data. Here, the information source is one chunk included in the company information identified on the screen 110. In this embodiment, by setting the information source included in the prompt to the identified chunk of company information, it is possible to request the generation of registered question data and registered answer data while satisfying the token restrictions in the large-scale language model 511.

[0147] The text data 12c indicates the format of the output data when the large-scale language model 511 outputs data in response to a prompt. Specifically, the text data 12c indicates that the format of the output data when the large-scale language model 511 outputs a set of registered question data and registered answer data generated by the large-scale language model 511 is to be JSON format data.

[0148] In this embodiment, by causing the large-scale language model 511 to output pairs of registered question data and registered answer data in JSON format, the information processing device 200 can automatically recognize pairs of registered question data and registered answer data included in the output data output from the large-scale language model 511. Therefore, in this embodiment, the information processing device 200 can generate a list of pairs of registered question data and registered answer data from the output data output from the large-scale language model 511, using chunks included in the identified internal company information as information sources.

[0149] Fig. 13 is a second diagram illustrating the prompt of the first embodiment. The prompt 13 shown in Fig. 13 is generated by the prompt generation unit 255 in step S915 of Fig. 9. In other words, the prompt 13 shown in Fig. 13 is a prompt requesting that the number of pairs of registered question data and registered answer data be set to a number corresponding to the setting information input on the screen 110.

[0150] 13 includes text data 13a and 13b. Text data 13a is text data that indicates to large-scale language model 511 that the role of large-scale language model 511 is to generate registered question data and registered answer data, and that a set number of pairs of registered question data and registered answer data should be selected from pairs that have already been generated. Text data 13b includes pairs of registered question data and registered answer data that have already been generated.

[0151] In this embodiment, the information processing device 200 outputs such a prompt to the large-scale language model 511, thereby being able to obtain a set number of pairs of registered question data and registered answer data from chunks contained in the identified internal company information.

[0152] The set of registered question data and registered answer data may be selected randomly by the large-scale language model 511 or may be selected according to conditions set by the administrator of the chat system 100.

[0153] Specifically, in this embodiment, for example, categories and themes that should be preferentially selected may be set by an administrator of the chat system 100. This setting may be performed, for example, on the screen 110, and the set content may be included as part of the setting information.

[0154] Next, a display example of a list of pairs of registered question data and registered answer data corresponding to output data output from large-scale language model 511 will be described with reference to Fig. 14. Fig. 14 is a second diagram showing a display example of the terminal device of the first embodiment.

[0155] A screen 140 shown in FIG. 14 is an example of a list screen displayed on the terminal device 700 in step S920 of FIG.

[0156] 13, the screen 140 includes a display area 141. In the example of Fig. 13, a list of registered question data is displayed in the display area 141. Specifically, the display area 141 includes display areas 141a, 141b, ..., in which registered question data generated by the large-scale language model 511 are displayed. Includes 141n.

[0157] In this embodiment, when a display area displaying registered question data is selected, screen 140 transitions to screen 150 (see FIG. 15) that displays the selected registered question data and registered answer data for this registered question data.

[0158] 14, a list of registered question data is displayed in display area 141. However, the present invention is not limited to this. Display area 141 may also display a list of pairs of registered question data and registered answer data.

[0159] In the following explanation, it is assumed that the display area 141a on the screen 140 displaying the registered question data "I want to know the rules for claiming travel expenses" is selected.

[0160] 15 is a third diagram showing a display example of the terminal device according to the first embodiment. A screen 150 shown in FIG. 15 is an example of an edit screen displayed on the terminal device 700 in step S1005 of FIG.

[0161] 15 includes display areas 151 and 152, and operation buttons 153 and 154. Display area 151 displays the registered question data selected on screen 140, which is a list screen. Specifically, display area 151 displays the registered question data displayed in display area 141a of screen 140.

[0162] Display area 152 displays the registered answer data corresponding to the registered question data displayed in display area 151. Operation button 153 is an operation button for ending the display of screen 150. Operation button 154 is an operation button for storing the question and answer information including the registered question data and registered answer data displayed in display areas 151 and 152 in question and answer storage unit 320.

[0163] In this embodiment, by displaying the registered question data and registered answer data generated by the large-scale language model 511 in correspondence with each other, the user of the terminal device 700 (the administrator of the chat system 100) can confirm the contents of the registered question data and registered answer data.

[0164] Furthermore, in the display area 152 of this embodiment, the registered question data displayed in the display area 151 and the registered answer data displayed in the display area 152 can be edited (corrected).

[0165] The registered question data and registered answer data displayed in the display areas 151 and 152 are text data. In this embodiment, the user of the terminal device 700 can modify the content and expression of the registered question data and registered answer data generated by the large-scale language model 511 by editing the registered question data and registered answer data displayed in the display areas 151 and 152, respectively.

[0166] In this embodiment, the registered answer data may include information other than answers to the registered question data.

[0167] Specifically, for example, the registered answer data may include, in addition to the answer corresponding to the registered question data, information indicating contact information for the user of the terminal device 600 in case of further inquiries. Furthermore, in addition to the answer corresponding to the registered question data, the registered answer data of this embodiment may include, in addition to the answer corresponding to the registered question data, a link for displaying information related to the registered question data and the content indicated by the registered answer data.

[0168] 15, information 152a indicating contact information and a link 152b are displayed in the display area 152. In this embodiment, the information 152a indicating contact information and the link 152b may be included in the output data output from the large-scale language model 511, or may be input by the user of the terminal device 700 in an editing operation.

[0169] In this embodiment, when the operation button 154 is operated on the screen 150, the registered question data displayed in the display area 151 and the registered answer data displayed in the display area 152 are linked and stored in the question and answer storage unit 320 via the information processing device 200.

[0170] In this embodiment, when registration question data is selected on screen 140, which is a list screen, screen 140 transitions to screen 150, but this is not limiting. For example, when registration question data is selected on screen 140, screen 150 may be displayed on screen 140 in a pop-up format or the like.

[0171] In this manner, in this embodiment, the administrator of the chat system 100 can check the contents of the registered question data and registered answer data generated by the large-scale language model 511, and can also edit them.

[0172] Therefore, according to this embodiment, it is possible to prevent inappropriate answer-to-question information from being stored in the question and answer storage unit 320, thereby improving the accuracy of the output answer data that is output in response to the input question data.

[0173] Next, a case where question data is input from the terminal device 600 will be described.

[0174] When input question data is input from the terminal device 600, the information processing device 200 of this embodiment causes the question and answering unit 260 to make a search request to the first storage device 300, and obtains the search results from the question and answer storage unit 320 from the first storage device 300. Then, the information processing device 200 outputs the registered question data obtained as the search results from the question and answer storage unit 320 to the terminal device 600 as output answer data.

[0175] 16 to 18, examples of the display of the terminal device 600 will be described below. Fig. 16 is a fourth diagram showing an example of the display of the terminal device of the first embodiment.

[0176] 16 is an example of a screen displayed on the terminal device 600. The screen 161A includes display areas 161a, 161b, and 161d. The display area 161a displays input question data input by the user of the terminal device 600. The display area 161b displays output answer data output from the information processing device 200 to the terminal device 600.

[0177] The output answer data displayed here is registered answer data associated with registered question data identified as a result of searching the question and answer storage unit 320 with the input question data.

[0178] The display area 161b includes text data 161b-1 and a link 161b-2. The text data 161b-1 includes information indicating contact information for the user of the terminal device 600 in case of further inquiries.

[0179] In this embodiment, the information indicating the contact point may be, for example, an email address or a telephone number, and may be stored in the question and answer storage unit 320 in association with the registered answer data.

[0180] In this case, when transmitting output answer data to the terminal device 600, the information processing device 200 may obtain an email address associated with the registered question data identified as a result of searching the question and answer storage unit 320 with the input question data, and transmit the email address to the terminal device 600 as part of the text data 161b-1.

[0181] The link 161b-2 is a link for displaying information related to the content indicated by the input question data or the output answer data.

[0182] The display area 161d displays a message notifying that the link 161b-2 displayed in the display area 161b has been sent to the email address of the user of the terminal device 600.

[0183] In this case, when receiving input question data from the terminal device 600, the information processing device 200 may receive the tenant ID and user ID of the tenant to which the user of the terminal device 600 belongs along with the input question data. The information processing device 200 may refer to the user information storage unit 420 and acquire an email address identified from the tenant ID and user ID received along with the input question data. Then, when transmitting output answer data to the terminal device 600, the information processing device 200 may transmit the link 161b-2 together with the output answer data to the identified email address, and then cause the terminal device 600 to display the display area 161d.

[0184] In this embodiment, by notifying the user of the terminal device 600 of contact information for further inquiries and information related to the question and answer as output answer data, it is possible to provide an answer that satisfies the user.

[0185] Furthermore, the information processing device 200 of this embodiment may send a notification to the contact information included in the text data 161b-1, indicating that an inquiry has been received from the user of the terminal device 600. By sending such a notification, it is possible to prompt the user who is the contact information when making an inquiry to make preparations for responding to further inquiries.

[0186] In the example of FIG. 16, by sending the link 161b-2 to the user's email address, the user can refer to information related to the inquiry even after finishing using the chat system 100.

[0187] Furthermore, in this embodiment, the document storage unit 410 may store conversation history information that associates input question data input from the terminal device 600 with output response data output in response to this input question data.

[0188] In this embodiment, by storing conversation history information in the conversation history storage unit 310 in this way, it is possible to utilize the history of past questions and answers.

[0189] 17 is a fifth diagram showing an example of a display on the terminal device 600 according to the first embodiment. A screen 171B shown in FIG.

[0190] The screen 161B includes display areas 161a and 161e. The display area 161e includes text data 161b-1 indicating contact information for the user of the terminal device 600 to use in case of further inquiries, and a message 161f.

[0191] Message 161f is a message indicating that a link to display information related to the user's inquiry has been sent to the user's email address.

[0192] In this way, in the example of FIG. 17, a link for displaying information related to the user's inquiry can be notified to the user separately from the output response data.

[0193] 18 is a sixth diagram illustrating an example of a display on the terminal device 600 according to the first embodiment. A screen 161C illustrated in FIG. 18 is an example of a screen displayed on the terminal device 600 when the user's email address is not stored in the user information storage unit 420 of the second storage device 400.

[0194] Screen 161C includes display areas 161a, 161b, 161g, 161h, and 161i. Display area 161g displays a message inquiring about the destination of link 161b-2. Display area 121h displays the email address entered by the user. Display area 121i displays a message notifying the user that link 121b-2 has been sent to the email address entered by the user.

[0195] In the example of FIG. 18, even if the user's email address is not stored in the user information storage unit 420, the link 121b-2 can be sent to the user if the user so desires.

[0196] 9 and 10, when the question and answer information is stored in the question and answer storage unit 320, the newly stored question and answer information may be used as learning data to be learned by the large-scale language model 511. In this way, the accuracy of the large-scale language model 511 can be improved.

[0197] (Second embodiment) The second embodiment will be described below with reference to the drawings. The second embodiment differs from the first embodiment in that the administrator of the chat system 100 inputs registration question data. In the following description of the second embodiment, differences from the first embodiment will be described, and components having the same functional configuration as those in the first embodiment will be assigned the same reference numerals as those used in the first embodiment, and their description will be omitted.

[0198] 19 is a first diagram showing a display example of the terminal device of the second embodiment. Fig. 19 is an example of a generated screen displayed on the terminal device 700 in step S901 of Fig. 9, for example.

[0199] A screen 110A shown in FIG. 19 includes an input area 111, a display area 112, operation buttons 113a, and an input field 117.

[0200] The input field 117 is an input field for inputting registered question data to be stored in the question and answer storage unit 320. The registered question data may be input by an administrator of the chat system 100 (a user of the terminal device 700).

[0201] Operation button 113a is an operation button for issuing an instruction to generate a set of registered question data and registered answer data.

[0202] When the operation button 113a is operated after the setting information is input and the in-house information is uploaded, the terminal device 700 sends an instruction to the information processing device 200 to generate registered response data using the uploaded in-house information in accordance with the setting information.

[0203] The operation of the information processing device 200 in this embodiment will be described below. Note that the operation of the information processing device 200 in this embodiment is the same as that in Fig. 9 except that the information that the information processing device 200 receives from the terminal device 700 in step S907 and the prompt that is generated in step S911 are different.

[0204] The information that the information processing device 200 of this embodiment receives from the terminal device 700 in step S907 is the setting information, registration question data, and identified in-house information input on the screen 110A of the terminal device 700. When the operation button 113a on the screen 110A of the terminal device 700 is operated, this information is transmitted to the information processing device 200.

[0205] The prompt generated by the information processing device 200 of this embodiment in step S911 is a prompt that requests the large-scale language model 511 to generate registered answer data corresponding to the input registered question data.

[0206] The following describes the prompts generated by the information processing device 200 of this embodiment. Fig. 20 is a first diagram illustrating the prompts of the second embodiment.

[0207] Prompt 20 shown in Figure 20 is a prompt that requests that registered answer data to the registered question data entered on screen 110A be output as JSON format data using chunks contained in the internal company information identified on screen 110A.

[0208] Prompt 20 includes text data 20a, 20b, and 20c. Text data 20a is text data that indicates to large-scale language model 511 that the role of large-scale language model 511 is to generate registered question data and registered answer data, and that registered question data for input registered question data is generated from chunks included in prompt 20 and output as data in JSON format.

[0209] The text data 20c indicates a format in which data in response to a prompt is output by the large-scale language model 511. The text data 20c may include the registration question data entered in the input field 117 on the screen 110A.

[0210] In this embodiment, by specifying in advance the internal company information serving as the information source and the registered question data, it is possible to generate registered answer data for the registered question data specified by the administrator of the chat system 100 in the large-scale language model 511. Therefore, according to this embodiment, it is possible to prevent question and answer information of content not intended by the administrator of the chat system 100 from being stored in the question and answer storage unit 320.

[0211] Furthermore, according to this embodiment, since the large-scale language model 511 generates the registration question data, the administrator of the chat system 100 does not need to prepare the registration question data, thereby reducing the burden on the administrator of the chat system 100.

[0212] Furthermore, when registered question data is input in advance on screen 110A as in this embodiment, the editing screen displayed on terminal device 700 in step S1005 of Fig. 10 may be displayed in a state where the registered answer data can be edited but editing of the registered question data is prohibited. In other words, in this embodiment, the editing screen displayed on terminal device 700 may be a screen where at least the registered answer data can be edited.

[0213] In this embodiment, a prompt may be output to the large-scale language model 511 requesting the generation of both registered answer data corresponding to the registered question data entered by the administrator of the chat system 100, and a pair of registered question data and registered answer data.

[0214] In this embodiment, instead of the registered question data, the administrator of the chat system 100 may specify the theme of the registered question data and registered answer data that must be generated.

[0215] A case where the administrator of the chat system 100 specifies a theme will be described below with reference to FIG.

[0216] Fig. 21 is a second diagram showing a display example of the terminal device of the second embodiment. A screen 110B shown in Fig. 21 includes an input area 111, a display area 112, operation buttons 113, and an input field 117A.

[0217] A theme for generating a set of registered question data and registered answer data is input into input field 117A.

[0218] In this embodiment, when the operation button 113 is operated on the screen 110B, the prompt generation unit 255 of the information processing device 200 generates a prompt requesting the generation of a pair of registered question data and registered answer data, including a pair of registered question data and registered answer data for the theme entered in the input field 117A.

[0219] Fig. 22 is a second diagram illustrating a prompt according to the second embodiment. Prompt 20A shown in Fig. 22 is a prompt requesting that registered question data and registered answer data be generated using chunks included in the in-house information identified on screen 110B so as to include a set of registered question data and registered answer data for the theme entered on screen 110B, and that the data be output as JSON format data.

[0220] Prompt 20A includes text data 22a, 12b, 22b, and 12c. Text data 22a is text data that indicates to large-scale language model 511 that the role of large-scale language model 511 is to generate registered question data and registered answer data, and that it will generate pairs of registered question data and registered answer data including pairs of registered question data and registered answer data for the theme included in prompt 20A from chunks included in prompt 20A, and output them as data in JSON format.

[0221] The text data 22b is text data that indicates the theme for generating a pair of registered question data and registered answer data. In other words, the text data 22b is information that indicates the theme of the pair of registered question data and registered answer data to be included in the output data.

[0222] In this embodiment, the administrator of the chat system 100 can specify the theme of the question and answer information to be stored in the question and answer storage unit 320, and have the large-scale language model 511 generate pairs of registered question data and registered answer data for the specified theme.

[0223] Therefore, according to this embodiment, the administrator of the chat system 100 can simply input a theme on the screen 110B to generate a pair of registered question data and registered answer data that includes a pair of registered question data and registered answer data related to this theme.

[0224] (Third embodiment) The third embodiment will be described below with reference to the drawings. The third embodiment differs from the first embodiment in that the administrator of the chat system 100 notifies the large-scale language model 511 of the role of the chat system 100. In the following description of the third embodiment, differences from the first embodiment will be described, and components having the same functional configuration as those in the first embodiment will be assigned the same reference numerals as those used in the first embodiment, and descriptions thereof will be omitted.

[0225] 23 is a first diagram showing a display example of the terminal device of the third embodiment. For example, Fig. 23 is an example of a generated screen displayed on the terminal device 700 in step S901 of Fig. 9.

[0226] The screen 110C shown in FIG. 23 includes an input area 111, a display area 112, operation buttons 113, and an input field 118.

[0227] The input field 118 is an input field for inputting information indicating the role of the chat system 100. The information indicating the role of the chat system 100 may be input by an administrator of the chat system 100.

[0228] Specifically, the information indicating the role of the chat system 100 may be, for example, information indicating the type of organization to which a user of the chat system 100 belongs, or information indicating the situation when a user of the chat system 100 inputs a question to the chat system 100. In other words, the information indicating the role of the chat system 100 is information indicating the role of the service provided by the information processing device 200 using the question and answer storage unit 320 in the input field 118.

[0229] In this embodiment, when the operation button 113 is operated after internal company information is identified and setting information and information indicating the role of the chat system 100 are entered, the terminal device 700 sends an instruction to the information processing device 200 to generate a set of registered question data and registered answer data.

[0230] The operation of the information processing device 200 in this embodiment will be described below. Note that the operation of the information processing device 200 in this embodiment is the same as that in Fig. 9 except that the information that the information processing device 200 receives from the terminal device 700 in step S907 and the prompt that is generated in step S911 are different.

[0231] The information that the information processing device 200 of this embodiment receives from the terminal device 700 in step S907 is the setting information input on the screen 110C of the terminal device 700, the information indicating the role of the chat system 100, and the identified in-house information. When the operation button 113 is operated on the screen 110C of the terminal device 700, these pieces of information are transmitted to the information processing device 200.

[0232] Furthermore, the prompt generated by the information processing device 200 of this embodiment in step S911 includes information indicating the role of the chat system 100.

[0233] The following describes prompts generated by the information processing device 200 of this embodiment. Fig. 24 is a diagram illustrating prompts of the third embodiment.

[0234] Prompt 24 shown in Figure 24 is a prompt that requests that a set of registered question data and registered answer data corresponding to the role of the chat system 100 entered on screen 110C be output as JSON format data using a chunk contained in the internal company information identified on screen 110C.

[0235] The prompt 24 includes text data 12a, 12b, 24a, and 12c. The text data 24a is text data that indicates the role of the chat system 100. In other words, the text data 24a includes the text data entered in the input field 118 on the screen 110C.

[0236] In the example of FIG. 24, the text data 24a indicates that the chat system 100 is installed on an in-house portal site and has a role of answering questions from employees.

[0237] In this embodiment, by including information indicating the role of the chat system 100 in the prompt input to the large-scale language model 511, the contents of the registered question data and registered answer data generated in the large-scale language model 511 can be made to suit the needs of users of the chat system 100. Therefore, according to this embodiment, the accuracy of the chat system 100 can be improved.

[0238] In this embodiment, the large-scale language model 511 may be requested to output categories corresponding to pairs of registered question data and registered answer data, and patterns obtained by paraphrasing the registered question data.

[0239] In other words, the information processing device 200 of this embodiment may include text data in the prompt output to the large-scale language model 511 requesting the output of a category corresponding to a pair of registered question data and registered answer data, and a pattern that rephrases the registered question data.

[0240] Fig. 25 is a diagram illustrating another prompt of the third embodiment. Prompt 25 shown in Fig. 25 is obtained by adding text data to prompt 24 described in Fig. 24, which requests output of categories corresponding to pairs of registered question data and registered answer data, and patterns that rephrase the registered question data.

[0241] Prompt 25 includes text data 25a, 12b, 24a, and 12c. Text data 25a includes text data requesting output of a category corresponding to a pair of registered question data and registered answer data, and a paraphrased version of the registered question data. In other words, text data 25a includes text data requesting output of a different version of the registered question data.

[0242] In response to the prompt, the large-scale language model 511 of the server device 500 outputs pairs of registered question data and registered answer data, categories corresponding to these pairs, and patterns obtained by paraphrasing the registered question data to the information processing device 200. The information processing device 200 causes the terminal device 700 to display a list of the registered question data and the registered answer data based on the output data acquired from the server device 500. Furthermore, when a pair of registered question data and registered answer data is selected on the terminal device 700, the information processing device 200 causes the terminal device 700 to display an editing screen including the pair of registered question data and registered answer data, categories, and patterns obtained by paraphrasing the registered question data.

[0243] 26 is a second diagram showing a display example of the terminal device of the third embodiment. A screen 150A shown in FIG. 26 is an example of an edit screen displayed on the terminal device 700 in step S1005 of FIG.

[0244] A screen 150A shown in FIG. 26 includes display areas 151 and 152, operation buttons 153 and 154, and display areas 155 and 156.

[0245] Display area 155 displays the categories of the registered question data and registered answer data displayed in display areas 151 and 152. In the example of Fig. 26, the registered question data and registered answer data displayed in display areas 151 and 152 are related to travel expense reimbursement, so display area 155 displays "travel expenses" as the category.

[0246] Display area 156 displays patterns of registered question data in which the registered question data displayed in display area 151 has been rephrased (paraphrased) so that the content of the question remains unchanged. In other words, a different pattern of the registered question data is displayed in display area 156. In the following description, the rephrased registered question data may be referred to as changed registered question data.

[0247] In the example of FIG. 26, three pieces of changed registered question data are displayed, but the number of changed registered question data may be any number and does not have to be three.

[0248] In this manner, in this embodiment, categories of pairs of registered question data and registered answer data, and patterns that rephrase the registered question data, are generated in the large-scale language model 511. Therefore, in this embodiment, the administrator of the chat system 100 does not need to refer to question and answer information to classify it into categories, or manually input patterns that change the wording of the registered question data, thereby further reducing the load on the chat system 100.

[0249] (Fourth embodiment) The fourth embodiment will be described below with reference to the drawings. The fourth embodiment differs from the first embodiment in that the information processing device 200 has a first storage device 300 and a second storage device 400. In the following description of the fourth embodiment, differences from the first embodiment will be described, and components having the same functional configuration as those in the first embodiment will be assigned the same reference numerals as those used in the first embodiment, and descriptions thereof will be omitted.

[0250] FIG. 27 is a diagram illustrating an example of a system configuration of a chat system according to the fourth embodiment.

[0251] The chat system 100A of this embodiment includes an information processing device 200A and a terminal device 700, and the information processing device 200A communicates with a server device 500 via the Internet. The information processing device 200A and the terminal device 700 are connected via a network. The information processing device 200A and the terminal device 700 may also be connected via a network other than the Internet. The chat system 100 of this embodiment may also include the server device 500.

[0252] In this embodiment, the information processing device 200A includes a question and answer storage unit 320 and a document storage unit 410. The question and answer storage unit 320 of this embodiment may store only question and answer information related to the organization in which the chat system 100A has been introduced. Furthermore, the document storage unit 410 of this embodiment may store only internal company information of the organization in which the chat system 100A has been introduced.

[0253] When the terminal device 700 identifies the in-house information to be used when generating a set of registered question data and registered answer data, the information processing device 200A stores the identified in-house information in the document storage unit 410 for each chunk.

[0254] Furthermore, when the information processing device 200A is instructed by the terminal device 700 to generate a set of registered question data and registered answer data, the information processing device 200A generates a prompt including a chunk included in the in-house information and text data requesting output of the set of registered question data and registered answer data that references the chunk in JSON format. The information processing device 200A then outputs the generated prompt to the server device 500.

[0255] In addition, when the information processing device 200 acquires output data output by the large-scale language model 511 provided in the server device 500, it generates a list of pairs of registered question data and registered answer data based on the output data and displays it on the terminal device 700.

[0256] When the administrator of the chat system 100A confirms the contents of the registered question data and registered answer data in the terminal device 700, the information processing device 200A stores the set of registered question data and registered answer data in the question and answer storage unit 320.

[0257] In this embodiment, by providing the information processing device 200A with the question and answer storage unit 320 and the document storage unit 410, it is possible to improve the confidentiality of the question and answer information stored in the question and answer storage unit 320 and the internal company information stored in the document storage unit 410.

[0258] Furthermore, in this embodiment, even if chat system 100A including information processing device 200A and terminal device 700 is built within a specific organization, it is possible to generate pairs of registered question data and registered answer data in large-scale language model 511. This reduces the burden on the administrator of chat system 100A.

[0259] The functions of the above-described embodiments can be realized by one or more processing circuits. Here, the term "processing circuit" in this specification includes a processor programmed to execute each function by software, such as a processor implemented by an electronic circuit, as well as devices such as an ASIC (Application Specific Integrated Circuit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), and conventional circuit modules designed to execute each of the above-described functions.

[0260] Additionally, the devices described in each embodiment represent only one of multiple computing environments for implementing the embodiments disclosed herein.

[0261] In one embodiment, information processing apparatus 200 includes multiple computing devices, such as a server cluster, configured to communicate with each other via any type of communication link, including a network, shared memory, etc., to perform the processes disclosed herein. Similarly, server apparatus 500 and information processing apparatus 200 may include multiple computing devices configured to communicate with each other.

[0262] Furthermore, the information processing device 200 can be configured to share the disclosed processing steps in various combinations. For example, a process performed by a predetermined unit can be executed by the information processing device 200. Similarly, the function of a predetermined unit can be executed by the information processing device 200. Furthermore, in the information processing device 200, each element may be integrated into a single server device or may be separated into multiple devices.

[0263] The information processing device 200 may be any device that has a communication function. The information processing device 200 may be, for example, a PJ (Projector), an IWB (Interactive White Board: a white board with an electronic blackboard function that allows mutual communication), an output device such as digital signage, a HUD (Head Up Display) device, industrial machinery, an imaging device, a sound collection device, medical equipment, a network home appliance, an automobile (Connected Car), a notebook PC (Personal Computer), a mobile phone, a smartphone, a tablet terminal, a game console, a PDA (Personal Digital Assistant), a digital camera, a wearable PC, a desktop PC, or the like.

[0264] For example, aspects of the present invention are as follows. <1> an input receiving unit that receives input of instructions for generating registered question data and registered answer data; a request unit that requests a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; an output unit that outputs the registered response data output from the large-scale language model in response to the request to a terminal device. <2> The request unit further requesting the generation of registration question data based on the information specified in the generation instruction; The output unit outputting to the terminal device a set of the registered question data and the registered answer data corresponding to the registered question data, the set being included in the output data output from the large-scale language model; <1> The information processing device described. <3> an editing unit that displays, on the terminal device, an editing screen that includes registered answer data included in at least the output data output from the large-scale language model and accepts edits to the registered answer data; <1> or <2> The information processing device described. <4> The request unit when the number of pairs of the registered question data and the registered answer data corresponding to the registered question data included in the output data output from the large-scale language model is greater than a predetermined number, requesting the large-scale language model to set the number of pairs of the registered question data and the registered answer data corresponding to the registered question data to the predetermined number; <2> or <3> The information processing device described. <5> The request unit a request for generating registered answer data that includes a portion of the specified information and corresponds to registered question data based on the portion of the specified information; <1> ~ <4> 10. The information processing device according to claim 9, wherein: <6> The request unit requesting that the output data be output in JSON format; <1> ~ <5> 10. The information processing device according to claim 9, wherein: <7> the request includes pre-entered registration question data; the request unit requests generation of registration response data corresponding to the input registration response data. <1> The information processing device described. <8> The request unit the information indicating a theme of the registered question data and a set of the registered question data and the corresponding registered answer data, and requests that the output data output from the large-scale language model include the registered question data and the set of the registered answer data corresponding to the theme; <1> ~ <7> 10. The information processing device according to claim 9, wherein: <9> The request is the registered question data and the registered answer data are stored in a question and answer storage unit, and <1> ~ <8> 10. The information processing device according to claim 9, wherein: <10> The request unit a request to output a category corresponding to a set of the registered question data and the registered answer data corresponding to the registered question data, and other registered question data in which the registered question data is expressed differently; <1> ~ <9> 10. The information processing device according to claim 9, wherein: <11> An information processing system including an information processing device and a terminal device that communicates with the information processing device, The information processing device includes: an input receiving unit that receives input of instructions for generating registered question data and registered answer data; a request unit that requests a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; an output unit that outputs, to the terminal device, registered answer data output from the large-scale language model in response to the prompt request; The terminal device An information processing system having a display control unit that displays a screen including the registered question data and the registered answer data. <12> An information processing method including an information processing device and a terminal device that communicates with the information processing device, The information processing device, Accepts input of instructions for generating registered question data and registered answer data, requesting a large-scale language model to generate registered answer data corresponding to the registered question data based on information specified in the generation instruction; outputting the registration response data output from the large-scale language model in response to the request to the terminal device; The terminal device, an information processing method for displaying a screen including the registered question data and the registered answer data; <13> Accepts input of an instruction to generate registered question data and registered answer data to be stored in a question and answer storage unit; requesting a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; a program for causing an information processing device to execute a process of outputting, to a terminal device, the registered response data output from the large-scale language model in response to the request;

[0265] Although the present invention has been described above based on the embodiments, the present invention is not limited to the requirements shown in the above embodiments. These requirements can be changed without departing from the spirit of the present invention, and can be appropriately determined depending on the application form. [Explanation of symbols]

[0266] 100, 100A chat system 200, 200A Information processing equipment 251 Input Reception Unit 252 Storage control unit 253 Setting reference section 254 Information Acquisition Department 255 Prompt Generation Unit 256 Request part 257 List Generation Unit 258 output section 259 Editorial Department 260 Question and answer section 300 First Storage Device 310 Conversation history memory unit 320 Question and answer storage section 400 Secondary Storage Device 410 Document Storage Unit 420 User information storage unit 500 Server device 511 Large-scale Language Models 600, 700 terminal equipment [Prior art documents] [Patent documents]

[0267] [Patent Document 1] Japanese Patent Publication No. 2023-181204

Claims

1. an input receiving unit that receives input of instructions for generating registered question data and registered answer data; a request unit that requests a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; an output unit that outputs the registered response data output from the large-scale language model in response to the request to a terminal device.

2. The request unit further requesting the generation of registration question data based on the information specified in the generation instruction; The output unit The information processing apparatus according to claim 1 , wherein a set of the registered question data and the registered answer data corresponding to the registered question data, which is included in the output data output from the large-scale language model, is output to the terminal device.

3. 3. The information processing apparatus according to claim 1, further comprising an editing unit that displays, on the terminal device, an editing screen that includes registered response data included in at least the output data output from the large-scale language model and accepts edits to the registered response data.

4. The request unit 3. The information processing device according to claim 2, wherein, when the number of pairs of the registered question data and the registered answer data corresponding to the registered question data included in the output data is greater than a predetermined number, the large-scale language model is requested to set the number of pairs of the registered question data and the registered answer data corresponding to the registered question data to the predetermined number.

5. The request unit 3. The information processing apparatus according to claim 2, wherein the information processing apparatus requests the generation of registered answer data that includes a portion of the specified information and corresponds to registered question data based on the portion of the specified information.

6. The request unit The information processing device according to claim 1 or 2, wherein the information processing device requests that the output data output from the large-scale language model be output in JSON format.

7. the request includes pre-entered registration question data; The information processing apparatus according to claim 1 , wherein the request unit requests the generation of registered response data corresponding to the input registered response data.

8. The request unit 3. The information processing device of claim 2, wherein the prompt includes information indicating a theme of the registered question data and a pair of the registered question data and the corresponding registered answer data, and requests that the output data include a pair of the registered question data and the corresponding registered answer data that is in line with the theme.

9. The request is 3. The information processing device according to claim 1, further comprising information indicating a role of a service using the registered question data and the registered answer data stored in a question and answer storage unit.

10. The request unit 3. The information processing apparatus according to claim 2, wherein the information processing apparatus requests output of a category corresponding to a set of the registered question data and the registered answer data corresponding to the registered question data, and other registered question data in which the registered question data is expressed differently.

11. An information processing system including an information processing device and a terminal device that communicates with the information processing device, The information processing device includes: an input receiving unit that receives input of instructions for generating registered question data and registered answer data; a request unit that requests a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; an output unit that outputs, to the terminal device, the registered response data output from the large-scale language model in response to the request; The terminal device An information processing system having a display control unit that displays a screen including the registered question data and the registered answer data.

12. An information processing method including an information processing device and a terminal device that communicates with the information processing device, The information processing device, Accepts input of instructions for generating registered question data and registered answer data, requesting a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; outputting, to the terminal device, the registered response data output from the large-scale language model in response to the request; The terminal device, an information processing method for displaying a screen including the registered question data and the registered answer data;

13. Accepts input of instructions for generating registered question data and registered answer data, requesting a large-scale language model to generate registered answer data corresponding to the registered question data based on the generation instruction; a program for causing an information processing device to execute a process of outputting, to a terminal device, the registered response data output from the large-scale language model in response to the request;

Images