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

Abstract

To present an answer based on information unique to a user.MEANS FOR SOLVING THE PROBLEM: An information processing device includes: an input reception unit that receives an input of question data; a prompt output unit that when the input of the question data is received, outputs a prompt to request a large-scale language model to generate answer data to the question data based on information unique to a user; and an information output unit that outputs as output answer data the answer data that is output from the large-scale language model based on the prompt that is output from the prompt output unit, to a terminal device.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, large-scale language models have become known that are trained with large amounts of data using models with parameters or variables that are 10 times more numerous than conventional general language models. In these large-scale language models, when a prompt containing a small amount of example data is input, the pre-trained model understands the patterns in the prompt and generates and outputs words or sentences corresponding to the input. Note that a prompt is an example of text data that expresses instructions or questions to the large-scale language model. Summary of the Invention [Problem to be solved by the invention]

[0003] The large-scale language models described above are trained based on information that is publicly available on the Internet, etc. For this reason, it has been difficult for the large-scale language models described above to present answers based on information that is not publicly available, such as information specific to users who use the large-scale language models.

[0004] The disclosed technology has been developed in light of the above circumstances, and aims to present answers based on information specific to the user. [Means for solving the problem]

[0005] The disclosed technology is an information processing device having an input receiving unit that receives input of question data, a prompt output unit that, upon receiving the input of the question data, outputs a prompt to a large-scale language model requesting the generation of answer data to the question data based on user-specific information, and an information output unit that outputs the answer data output from the large-scale language model based on the prompt output from the prompt output unit to a terminal device as output answer data. [Effects of the Invention]

[0006] Answers can be provided based on user-specific information. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 illustrates an example of a system configuration of a chat system according to a first embodiment. [Figure 2] FIG. 1 illustrates an example of a hardware configuration of an information processing apparatus according to a first embodiment. [Figure 3] FIG. 2 illustrates an example of a hardware configuration of a terminal device according to the first embodiment. [Figure 4] FIG. 2 illustrates an example of a conversation history storage unit according to the first embodiment; [Figure 5] FIG. 2 illustrates an example of a question and answer storage unit according to the first embodiment; [Figure 6] FIG. 2 illustrates an example of a document storage unit according to the first embodiment; [Figure 7] FIG. 2 illustrates an example of a user information storage unit according to the first embodiment; [Figure 8] FIG. 2 is a diagram illustrating the functions of each device in the chat system of the first embodiment. [Figure 9] FIG. 3 is a sequence diagram illustrating the operation of the chat system according to the first embodiment. [Figure 10] FIG. 10 is a first diagram illustrating an example of a prompt according to the first embodiment. [Figure 11] FIG. 10 is a second diagram illustrating an example of a prompt according to the first embodiment. [Figure 12] FIG. 10 is a first diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 13] FIG. 10 is a second diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 14] FIG. 10 is a third diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 15] FIG. 10 is a fourth diagram illustrating a display example of the terminal device according to the first embodiment. [Figure 16] FIG. 10 is a sequence diagram illustrating the operation of a chat system according to a second embodiment. [Figure 17] FIG. 10 illustrates an example of a prompt according to the second embodiment. [Figure 18] FIG. 11 is a sequence diagram illustrating the operation of a chat system according to a third embodiment. [Figure 19] FIG. 11 is a diagram illustrating a display example of a terminal device according to the third embodiment. [Figure 20] FIG. 10 is a sequence diagram illustrating the operation of a chat system according to a fourth embodiment. [Figure 21] FIG. 13 is a diagram illustrating a display example of a terminal device according to the fourth embodiment. [Figure 22] FIG. 13 is a diagram illustrating an example of a prompt according to the fourth embodiment. [Figure 23] FIG. 13 is a sequence diagram illustrating the operation of a chat system according to a fifth embodiment. [Figure 24] FIG. 13 is a first diagram illustrating a display example of a terminal device according to the fifth embodiment. [Figure 25] FIG. 13 is a second diagram illustrating an example of a display on the terminal device according to the fifth embodiment. [Figure 26] FIG. 20 is a third diagram illustrating a display example of the terminal device according to the sixth embodiment. [Figure 27] FIG. 13 is a sequence diagram illustrating the operation of a chat system according to a sixth embodiment. [Figure 28] FIG. 13 is a first diagram illustrating an example of a prompt according to the sixth embodiment. [Figure 29] FIG. 20 is a second diagram illustrating an example of a prompt according to the sixth embodiment. [Figure 30]FIG. 20 is a third diagram illustrating an example of a prompt according to the sixth embodiment. [Figure 31] FIG. 13 is a diagram illustrating the functions of each device in the chat system of the seventh embodiment. [Figure 32] FIG. 19 is a diagram illustrating an example of a document storage unit according to the seventh embodiment. [Figure 33] 13 illustrates an example of a chunk storage unit according to the seventh embodiment. [Figure 34] FIG. 20 illustrates an example of a scenario storage unit according to the seventh embodiment. [Figure 35] FIG. 13 is a first sequence diagram illustrating the operation of the chat system according to the seventh embodiment. [Figure 36] FIG. 20 is a second sequence diagram illustrating the operation of the chat system according to the seventh embodiment. [Figure 37] FIG. 13 is a first diagram illustrating a display example of a terminal device according to the seventh embodiment. [Figure 38] FIG. 22 is a second diagram illustrating an example of a display on the terminal device according to the seventh embodiment. [Figure 39] FIG. 13 is a third diagram illustrating a display example of the terminal device according to the seventh embodiment. [Figure 40] FIG. 13 is a third sequence diagram illustrating the operation of the chat system according to the seventh embodiment. [Figure 41] FIG. 13 is a fourth diagram illustrating a display example of the terminal device according to the seventh 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 according to the first embodiment.

[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 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. Note that 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] 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.

[0015] 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.

[0016] 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.

[0017] 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.

[0018] The server device 500 has a large-scale language model storage unit 510 and realizes a generative AI (Artificial Intelligence). The generative AI generates text data according to an input prompt, such as ChatGPT, and outputs the text data to the information processing device 200. Note that the prompt in this embodiment is text data for expressing instructions or questions to the generative AI realized by the server device 500. Details of the large-scale language model will be described later.

[0019] The terminal device 600 is mainly used by a user who uses the chat system 100. The terminal device 700 is mainly used by an administrator who manages the chat system 100. Each of the terminal devices 600 and 700 may be, for example, a smartphone or a tablet terminal.

[0020] In the chat system 100 of this embodiment, when a user inputs a question to the terminal device 600, the terminal device 600 transmits question data indicating the question to the information processing device 200.

[0021] When the information processing device 200 receives input of question data, it acquires, from the conversation history storage unit 310 of the first storage device 300, conversation history information for a certain period of time before the question was input to the terminal device 600. Then, the information processing device 200 uses the input question and the acquired conversation history information to output to the server device 500 a prompt requesting the generation of a query (search key) for searching the document storage unit 410 of the second storage device 400.

[0022] The server device 500 inputs a prompt requesting the generation of a query to the large-scale language model stored in the large-scale language model storage unit 510, and transmits output data output from the large-scale language model to the information processing device 200. The information processing device 200 searches the document storage unit 410 of the second storage device 400 using the output data received from the server device 500 as a query, and obtains in-house information as a search result.

[0023] Then, the information processing device 200 outputs to the server device 500 a prompt requesting the server device 500 to generate an answer to the question input from the terminal device 600, using the in-house information acquired as the search result as an information source.

[0024] The server device 500 inputs a prompt requesting generation of an answer to the large-scale language model stored in the large-scale language model storage unit 510, and transmits output data output from the large-scale language model to the information processing device 200. The information processing device 200 transmits the output data received from the server device 500 to the terminal device 600 as answer data.

[0025] The terminal device 600 displays the answer data received from the information processing device 200 as the answer to the input question.

[0026] In this embodiment, by doing so, it is possible to have the large-scale language model generate answers to questions based on user-specific information that is not publicly available, thereby providing responses based on user-specific information.

[0027] In addition, in this embodiment, the information source when generating an answer is internal company information obtained by searching the document storage unit 410 based on the conversation history information before the question was entered. Therefore, according to this embodiment, the internal company information to be used as the information source can be narrowed down and transmitted to the server device 500, thereby improving the accuracy of the answer while satisfying the token restrictions in the large-scale language model.

[0028] 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.

[0029] In addition, in FIG. 1, the information processing device 200 may be realized by one or more information processing devices, or may be realized by a plurality of information processing devices.

[0030] Furthermore, the chat system 100 only needs to include the information processing device 200 and the terminal device 600, and other devices may not be included in the chat system 100.

[0031] Next, an example of the hardware configuration of each device included in the chat system 100 of this embodiment will be described.

[0032] 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.

[0033] In this embodiment, the terminal devices 600 and 700 are smartphones, and in FIG. 3, the hardware configuration of the terminal device 600 will be described as an example.

[0034] Fig. 2 is a diagram showing an example of the hardware configuration of an information processing apparatus according to the first embodiment. The information processing apparatus 200 shown in Fig. 2 is constructed by a computer (information processing apparatus), 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.

[0035] 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.

[0036] 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.

[0037] 3 is a diagram showing an example of the hardware configuration of a terminal device 600 according to the first embodiment. The terminal device 600 according to the present embodiment includes a CPU 601, a ROM 602, a RAM 603, an EEPROM 604, a CMOS sensor 605, an image sensor I / F 606, an acceleration / direction sensor 607, a media I / F 609, and a GPS receiver 611.

[0038] Of these, the CPU 601 is an arithmetic processing unit that controls the overall operation of the terminal device 600. The ROM 602 stores the CPU 601 and programs used to drive the CPU 601, such as the IPL. The RAM 603 is used as a work area for the CPU 601. The EEPROM 604 reads or writes various data, such as smartphone programs, under the control of the CPU 601. The ROM 602, RAM 603, and EEPROM 604 are examples of storage devices of the terminal device 600.

[0039] The CMOS (Complementary Metal Oxide Semiconductor) sensor 605 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 601. Note that instead of a CMOS sensor, an imaging means such as a CCD (Charge Coupled Device) sensor may also be used.

[0040] The imaging element I / F 606 is a circuit that controls the driving of the CMOS sensor 605. The acceleration / azimuth sensor 607 is a variety of sensors, such as an electronic magnetic compass that detects geomagnetism, a gyrocompass, and an acceleration sensor. The media I / F 609 controls the reading and writing (storage) of data from and to a recording medium 608, such as a flash memory. The GPS receiver 611 receives GPS signals from GPS satellites.

[0041] The terminal device 600 also includes a long-distance communication circuit 612, an antenna 612a of the long-distance communication circuit 612, a CMOS sensor 613, an image sensor I / F 614, a microphone 615, a speaker 616, an audio input / output I / F 617, a display 618, an external device connection I / F (Interface) 619, a short-distance communication circuit 620, an antenna 620a of the short-distance communication circuit 620, and a touch panel 621.

[0042] Of these, the long-distance communication circuit 612 is a circuit that communicates with other devices via a communication network. The CMOS sensor 613 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 601. The image sensor I / F 614 is a circuit that controls the driving of the CMOS sensor 613. The microphone 615 is a built-in circuit that converts sound into an electrical signal. The speaker 616 is a built-in circuit that converts an electrical signal into physical vibrations to produce sound such as music or voice. The audio input / output I / F 617 is a circuit that processes the input and output of audio signals between the microphone 615 and the speaker 616 under the control of the CPU 601.

[0043] The display 618 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 619 is an interface for connecting various external devices. The short-range communication circuit 620 is a communication circuit such as NFC (Near Field Communication) or Bluetooth (registered trademark). The touch panel 621 is a type of input means that allows a user to operate the terminal device 600 by pressing the display 618. The display 618 is an example of a display unit that the terminal device 600 has.

[0044] 3 is an example of the terminal device 600, and the hardware configuration of the terminal device 600 is not limited to this. The hardware configuration of the terminal device 600 may be the configuration shown in FIG.

[0045] 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 of the first embodiment.

[0046] 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.

[0047] 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 a question was input to the terminal device 600 or the date and time when an answer 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.

[0048] The value of the item "Question" indicates data (question data) input as a question sentence to the terminal device 600. The value of the item "Answer" indicates data (answer data) output as an answer sentence to the terminal device 600.

[0049] In the following description, data output as an answer to the terminal device 600 will be referred to as output answer data. Therefore, the conversation history information of 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 question data was input or the date and time when the output answer data was output, the question data, and the output answer data.

[0050] 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.

[0051] 5 is a diagram illustrating an example of the question and answer storage unit 320 according to the first embodiment. The question and answer information stored in the question and answer storage unit 320 according to this embodiment may be provided for each tenant and managed by each tenant. The example in FIG. 5 illustrates the question and answer information managed by the tenant identified by the tenant ID "11."

[0052] The question and answer information includes information items such as a tenant ID, a question and answer ID, an update date and time, a question, and an answer, 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.

[0053] 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 question data (text data) that indicates the question, and the value of the item "Answer" is answer data (text data) that indicates the answer. The values ​​of the items "Question" and "Answer" are stored in advance by the administrator of each tenant, etc.

[0054] In the following description, the question data stored in the question and answer storage unit 320 by the administrator of each tenant will be referred to as registered question data, and the answer data stored in the question and answer storage unit 320 by the administrator of each tenant will be referred to as registered answer data.

[0055] Fig. 6 is a diagram illustrating an example of a document storage unit according to the first embodiment. In the document storage unit 410 according to 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."

[0056] 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.

[0057] 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.

[0058] 7 is a diagram illustrating an example of the user information storage unit 420 according to the first embodiment. The user information stored in the user information storage unit 420 according to the present embodiment is in-house information stored for each tenant, and may be managed for each tenant.

[0059] 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. Note that the user information may include information that indicates the attributes of the user. The information that indicates the attributes of the user may be, for example, information that indicates the department to which the user belongs.

[0060] In addition, 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.

[0061] 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.

[0062] 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 search request unit 252, a conversation history acquisition unit 253, an information acquisition unit 254, a prompt generation unit 255, a prompt output unit 256, and an information output unit 257.

[0063] The input receiving unit 251 receives various inputs to the information processing device 200. Specifically, the input receiving unit 251 receives input of question data from the terminal device 600.

[0064] The search request unit 252 issues a search request to the first storage device 300 or the second storage device 400. Specifically, the search request unit 252 sends a search request for the conversation history storage unit 310 using the question data as a search key to the first storage device 300. The search request unit 252 also sends a search request for the document storage unit 410 based on a query generated by the server device 500 to the second storage device 400.

[0065] The conversation history acquisition unit 253 acquires, from the first storage device 300, conversation history information that is a search result of the conversation history storage unit 310 in response to a search request from the search request unit 252.

[0066] The information acquisition unit 254 acquires from the second storage device 400 the in-house information that is the search result of the document storage unit 410 in response to the search request made by the search request unit 252 .

[0067] The prompt generating unit 255 generates a prompt to be output to the server device 500. For example, when information to be included in the prompt is acquired, the prompt generating unit 255 generates a prompt in a predetermined format that includes the acquired information.

[0068] The prompt output unit 256 outputs the prompt generated by the prompt generation unit 255 to the server device 500. The information output unit 257 outputs various types of information from the information processing device 200.

[0069] Next, the functions of the first storage device 300 will be described. 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 conversation history storage unit 310 or 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. When 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 question data and the output answer data in the conversation history storage unit 310. 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 330 searches the conversation history storage unit 310 and 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. When an administrator of the chat system 100 or the like performs an operation to upload company information, the storage unit 440 stores the new company information in the document storage unit 410.

[0070] 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.

[0071] 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.

[0072] 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.

[0073] 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.

[0074] 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.

[0075] Next, the operation of the chat system 100 of this embodiment will be described with reference to Fig. 9. Fig. 9 is a sequence diagram illustrating the operation of the chat system of the first embodiment.

[0076] In the chat system 100, when the terminal device 600 receives a question input via the input receiving unit 640, the terminal device 600 transmits question data to the information processing device 200 (step S901). Note that the transmitted question data may be assigned the tenant ID of the tenant to which the user of the terminal device 600 belongs.

[0077] When the information processing device 200 receives input of question data via the input receiving unit 251, the search request unit 252 sends a search request including the tenant ID to the first storage device 300 (step S902).

[0078] When the first storage device 300 receives this search request, the search unit 330 searches the conversation history storage unit 310 (step S903).

[0079] Here, we will explain the processing of the search unit 330. The search unit 330 identifies conversation history information that includes the tenant ID included in the search request from the conversation history information stored in the conversation history storage unit 310, and obtains, from the identified conversation history information, conversation history information whose content is related to the question indicated by the question data as a search result.

[0080] Specifically, the search unit 330 identifies conversation history information that includes the tenant ID included in the search request, and from the identified conversation history information, obtains the conversation history information that continued immediately before the question data was input in step S901.

[0081] The conversation history information that continued immediately before the question data was input in step S901 is conversation history information for which the time between the time the question data was input and the time the output answer data was output immediately before the input of this question data is within a certain period of time. In this embodiment, if the next question data is input within a certain period of time after the output answer data was output, the series of questions and answers is considered to be responses to a common inquiry item.

[0082] In this embodiment, by acquiring conversation history information in this manner, conversation history information whose contents are related to the question indicated by the question data can be acquired.

[0083] After searching the conversation history storage unit 310, the search unit 330 of the first storage device 300 transmits the search results to the information processing device 200, and the information processing device 200 acquires the conversation history information via the conversation history acquisition unit 253 (step S904).

[0084] Here, when conversation history information is acquired from the first storage device 300 as a search result, the information processing device 200 performs the following processes of steps S905 to S907.

[0085] Note that conversation history information is acquired when, before the question data input in step S901 is input, question data related to the question indicated by the question data input in step S901 is input and corresponding output answer data is output. In this case, conversation history information including this question data and output answer data is stored in conversation history storage unit 310.

[0086] Furthermore, if the information processing device 200 does not acquire conversation history information from the first storage device 300 as a search result, the information processing device 200 proceeds to step S908. A case in which conversation history information is not acquired is when the question data input in step S901 indicates the first question in the inquiry.

[0087] The information processing device 200 causes the prompt generation unit 255 to generate a prompt requesting the generation of a query for searching the document storage unit 410 (step S905). At this time, the prompt generation unit 255 generates a prompt that includes the question data input in step S901 and the conversation history information acquired by the conversation history acquisition unit 253, and requests the generation of a query based on the question data and the conversation history information.

[0088] Next, the information processing device 200 causes the prompt output unit 256 to transmit a prompt requesting the server device 500 to generate a query (step S906).

[0089] The server device 500 inputs the prompt to the large-scale language model 511 via the input unit 520, and transmits the output data output from the large-scale language model 511 to the information processing device 200 as a query via the output unit 530 (step S907).

[0090] The information processing device 200, via the search request unit 252, sends a search request to the document storage unit 410 of the second storage device 400 for internal company information that can serve as an information source when generating output response data to the question data (step S908).

[0091] Specifically, if the search request unit 252 has acquired a query in step S907, it transmits a search request based on the query to the second storage device 400. If the search request unit 252 has not acquired a query, that is, if the search request unit 252 has not acquired conversation history information in step S904, it transmits a search request including the question data input in step S901 to the second storage device 400.

[0092] When the second storage device 400 receives the search request, the search unit 430 searches the document storage unit 410 (step S909), and transmits the in-house information of the search result to the information processing device 200 (step S910).

[0093] When the information processing device 200 acquires the internal company information from the search results using the information acquisition unit 254, the prompt generation unit 255 generates a prompt requesting the generation of an answer to the question data using the internal company information from the search results as an information source (step S911), and transmits the prompt from the server device 500 (step S912).

[0094] When the server device 500 receives this prompt, the input unit 520 inputs the prompt to the large-scale language model 511, and the output unit 530 transmits the output data output from the large-scale language model 511 to the information processing device 200 as answer output data (step S913).

[0095] When the information processing device 200 acquires the output response data, the information output unit 257 outputs the acquired output response data to the terminal device 600 (step S914). The terminal device 600 causes the display control unit 650 to display the output response data on the display 618 (step S915). At this time, the display control unit 650 causes the display 618 to display the output response data in association with the question data.

[0096] Furthermore, the information processing device 200 causes the information output unit 257 to transmit to the first storage device 300 a request to save conversation history information including the question data accepted in step S901 and the output answer data output to the terminal device 600 in step S912 (step S916). Here, the request to save conversation history information includes the tenant ID of the tenant to which the user of the terminal device 600 belongs.

[0097] In the first storage device 300, the storage unit 340 stores the new conversation history information in the conversation history storage unit 310 (step S917). Specifically, the storage unit 340 assigns a new conversation ID and date and time to the question data, output answer data, and tenant ID acquired from the information processing device 200, and stores them as conversation history information in the conversation history storage unit 310. In this embodiment, the conversation history storage unit 310 is updated in this way.

[0098] Next, examples of prompts output from information processing device 200 to server device 500 will be described with reference to FIGS.

[0099] 10 is a first diagram illustrating an example of a prompt in the first embodiment. Prompt 10 is an example of a prompt requesting generation of a query, which is transmitted from information processing device 200 to server device 500 in step S905 in FIG.

[0100] In the example of Figure 10, the question data input to the information processing device 200 in step S901 of Figure 9 is "Who should I contact for inquiries?", and the conversation history information acquired in step S904 includes the question "How do I record my attendance?" and the answer "The way to record your attendance is XXX."

[0101] The prompt 10 includes text data 10a, 10b, and 10c. The text data 10a indicates an inquiry from the user of the terminal device 600 and a request to generate a query for searching the document storage unit 410 based on the conversation history.

[0102] Furthermore, the text data 10b indicates the content of an inquiry from the user of the terminal device 600. Specifically, the text data 10b includes the question data "Who should I inquire about?" input of which was accepted by the information processing device 200 in step S901. Furthermore, the text data 10c indicates a question and answer included in the conversation history information acquired in step S904. Specifically, the text data 10c includes the question "How do I record my attendance?" and the answer "The way to record attendance is XXX." included in the conversation history information acquired by the information processing device 200 in step S904.

[0103] In this embodiment, by inputting such a prompt 10 into the large-scale language model 511 of the server device 500, the server device 500 can generate a query for searching the second storage device 400 from the question indicated by the input question data and conversation history information including content related to the question indicated by the question data.

[0104] Furthermore, in this embodiment, by searching the document storage unit 410 of the second storage device 400 with such a query, the information source when generating output answer data for question data can be internal company information that is likely to contain the answer to the question indicated by the question data. Therefore, according to this embodiment, the accuracy of the answer can be improved.

[0105] 11 is a second diagram showing an example of a prompt in the first embodiment. Prompt 11 is an example of a prompt that requests generation of output response data, which is transmitted from information processing device 200 to server device 500 in step S910 of FIG.

[0106] The prompt 11 includes text data 11a, 11b, and 11b. The text data 11a is text data that requests the generation of output answer data to a question indicated by question data, based on an information source.

[0107] The text data 11b is an information source. The information source indicated by the text data 11b is the company information acquired from the second storage device 400 in step S909 of Fig. 9. In other words, the information source indicated by the text data 11b is the company information acquired as a search result by searching the document storage unit 410 with a query generated from the inquiry content of the user of the terminal device 600 and the conversation history information.

[0108] In this embodiment, by inputting a prompt 11 containing internal company information as an information source into the large-scale language model 511 of the server device 500, an answer based on tenant-specific information can be obtained from the large-scale language model 511.

[0109] Furthermore, according to this embodiment, internal company information serving as an information source is obtained by a search based on a query including conversation history information. Therefore, according to this embodiment, an answer to a user's inquiry can be obtained from the large-scale language model 511 using internal company information that is likely to contain content related to the content of the user's inquiry as an information source, thereby improving the accuracy of the answer.

[0110] In addition, in this embodiment, a portion of the internal company information stored in the document storage unit 410 is extracted and included in the prompt, so that it is possible to obtain an answer that uses internal company information as a source of information while satisfying the limit on the number of tokens that can be given to the large-scale language model 511 as a prompt.

[0111] Next, examples of the display of the terminal device 600 will be described with reference to Fig. 12 to Fig. 15. Fig. 12 is a first diagram showing an example of the display of the terminal device of the first embodiment.

[0112] A screen 121 shown in FIG. 12 is an example of a screen displayed on the terminal device 600 in step S913 of FIG.

[0113] The screen 121 includes display areas 121a and 121b. The display area 121a displays question data input by the user of the terminal device 600. The display area 121b displays answer data output from the chat system 100, as well as text data 121b-1 and a link 121b-2.

[0114] The text data 121b-1 includes information indicating contact information for the user of the terminal device 600 in case of further inquiries. The information indicating contact information may be, for example, an email address or a telephone number. This contact information may be stored in the user information storage unit 420, for example. Specifically, the user information storage unit 420 may store information indicating the work responsibilities of each user in association with the user ID.

[0115] When transmitting the output response data to the terminal device 600, the information processing device 200 may obtain the email address of the user in charge of the business corresponding to the question indicated by the question data, and transmit the email address to the terminal device 600 together with the output response data.

[0116] The link 121b-2 is a link for displaying information related to the content indicated by the question data and the output response data.

[0117] 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 along with the output answer data, it is possible to provide the user with an answer that will satisfy them.

[0118] Furthermore, the information processing device 200 of this embodiment may send a notification to the contact information included in the text data 121b-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.

[0119] 13 is a second diagram showing a display example of the terminal device of the first embodiment. A screen 121A shown in FIG. 13 is another example of the screen displayed on the terminal device 600 in step S913 of FIG.

[0120] The screen 121A includes display areas 121a, 121b, and 121d. In the display area 121d, a message is displayed notifying that the link 121b-2 displayed in the display area 121b has been sent to the email address of the user of the terminal device 600.

[0121] In this case, the user ID of the user of the terminal device 600 may be input to the information processing device 200 together with the question data, and 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 the user ID. Then, when transmitting output response data to the terminal device 600, the information processing device 200 may transmit the link 121b-2 together with the output response data to the identified email address, and then cause the terminal device 600 to display the display area 121d.

[0122] In the example of Figure 13, by sending link 121b-2 to the user's email address to display information related to the user's inquiry, the user can refer to the information related to the inquiry even after finishing using the chat system 100.

[0123] 14 is a third diagram showing a display example of the terminal device of the first embodiment. Screen 121B shown in Fig. 14 is another example of the screen displayed on terminal device 600 in step S913 of Fig. 9.

[0124] Screen 121B includes display areas 121a and 121e. Display area 121e displays text data 121b-1 and a message 121f. Message 121f is a message indicating that a link for displaying information related to the user's inquiry has been sent to the user's email address.

[0125] In the example of FIG. 14, a link for displaying information related to the user's inquiry can be notified to the user separately from the output response data.

[0126] Fig. 15 is a fourth diagram showing a display example of the terminal device of the first embodiment. Screen 121C shown in Fig. 15 is another example of a screen displayed on terminal device 600 in step S913 of Fig. 9. Screen 121C may be displayed when the user's email address is not stored in user information storage unit 420 of second storage device 400.

[0127] Screen 121B includes display areas 121a, 121b, 121g, 121h, and 121i. Display area 121g displays a message inquiring about the destination of link 121b-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.

[0128] In the example of FIG. 15, 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.

[0129] (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 server device 500 corrects output response data generated by the large-scale language model 511. 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 description of the first embodiment, and descriptions thereof will be omitted.

[0130] Fig. 16 is a sequence diagram illustrating the operation of the chat system of the second embodiment. The processing from step S1601 to step S1613 in Fig. 16 is the same as the processing from step S901 to step S913 in Fig. 9, so the description will be omitted.

[0131] When the information processing device 200 acquires the output answer data from the server device 500, the prompt generation unit 255 generates a prompt requesting correction of the output answer data (step S1614) and transmits the generated prompt to the server device 500 (step S1615). When the server device 500 receives the prompt requesting correction, the input unit 520 inputs the prompt to the large-scale language model 511, and the output unit 530 transmits the output data output from the large-scale language model 511 to the information processing device 200 as correction result data (step S1616).

[0132] Note that, if the output answer data output in step S1613 is incorrect as a result of the correction, server device 500 may transmit correct output answer data as a result of the correction to information processing device 200. Also, if the output answer data output in step S1613 is correct, server device 500 may transmit the output answer data output in step S1613 to information processing device 200 again as a result of the correction.

[0133] Subsequently, the information processing device 200 transmits the correction result data received from the server device 500 to the terminal device 600 as output response data to the question data (step S1617).

[0134] The processing from step S1617 to step S1620 is the same as the processing from step S914 to step S917 in FIG. 9, and therefore a description thereof will be omitted.

[0135] Next, the prompt generated in step S1614 will be described with reference to Fig. 17. Fig. 17 is a diagram showing an example of a prompt according to the second embodiment.

[0136] The prompt 17 shown in FIG. 17 is a prompt requesting correction of the output response data output by the large-scale language model 511.

[0137] The prompt 17 includes text data 17a, 17b, 17c, and 17d. The text data 17a is text data indicating the content of a request for correction of the output answer data output from the large-scale language model 511. The text data 17b indicates the content of an inquiry from the user of the terminal device 600. Specifically, the text data 17b includes the question data "What kind of plans are available?" inputted and accepted by the information processing device 200 in step S1601.

[0138] Furthermore, the text data 17c includes the output response data acquired by the information processing device 200 in step S1613 of Fig. 16. Specifically, the text data 17c includes the output response data "There are three plans as follows: 1. Started plan, 2. Standard plan, and 3. Enterprise plan" acquired by the information processing device 200 in step S1613.

[0139] 16. The text data 17d indicates in-house information serving as an information source, which is acquired by the information processing device 200 in step S1610 of FIG.

[0140] In this embodiment, by having the large-scale language model 511 correct the output response data it has output to check for errors, it is possible to prevent incorrect responses from being output to the user of the terminal device 600, thereby improving the accuracy of the responses.

[0141] (Third embodiment) A third embodiment will be described below with reference to the drawings. The third embodiment differs from the first embodiment in that it is possible to edit in-house information stored in the document storage unit 410. 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 description of the first embodiment, and descriptions thereof will be omitted.

[0142] Fig. 18 is a sequence diagram illustrating the operation of the chat system of the third embodiment. Fig. 18 shows the operation of the chat system 100 when an administrator of the internal company information edits the internal company information stored in the document storage unit 410.

[0143] In the chat system 100, the terminal device 700 transmits a display request for the management screen of the chat system 100 to the information processing device 200 (step S1801). The information processing device 200 accepts this display request and transmits an instruction to display the management screen to the terminal device 700 (step S1802).

[0144] The terminal device 700 causes the display control unit 750 to display the management screen (step S1803). Note that the management screen of this embodiment may include operation buttons for transitioning to screens for editing, adding (uploading), and deleting in-house information, for example.

[0145] In step S1803 of FIG. 18, it is assumed that an operation to edit in-house information has been performed on the management screen.

[0146] When the terminal device 700 accepts an operation to edit the in-house information, it transmits a request to acquire a list of in-house information of the tenant to which the user of the terminal device 700 belongs to the information processing device 200 (step S1804). This request to acquire the list includes the tenant ID of the tenant to which the user of the terminal device 700 belongs.

[0147] When the information processing device 200 receives the request to acquire the list, the information acquisition unit 254 transmits the request to acquire the list including the tenant ID to the second storage device 400 (step S1805).

[0148] The second storage device 400 accepts the request to acquire the list, acquires a list of in-house information associated with the tenant ID included in the request to acquire the list from among the in-house information stored in the document storage unit 410 (step S1806), and transmits the list to the information processing device 200 (step S1807). Upon acquiring the list, the information processing device 200 transmits an instruction to display the acquired list to the terminal device 700 (step S1808). Note that the list of in-house information in this embodiment may be a list of chunks stored in the document storage unit 410.

[0149] The terminal device 700 accepts the display instruction and displays a list screen of in-house information (step S1809). Subsequently, when the terminal device 700 accepts an operation to select in-house information to be edited from the list screen (step S1810), it transmits an edit request for the selected in-house information to the information processing device 200 (step S1811). Note that the edit request may include identification information that identifies the selected in-house information (chunk).

[0150] When the information processing device 200 accepts the edit request, it transmits to the terminal device 700 an instruction to display an edit screen for the selected in-house information (step S1812).

[0151] The terminal device 700 receives the display instruction and displays the edit screen for the selected in-house information. (Step S1813), and accepts editing operations (Step S1814).

[0152] When the editing operation is completed, the terminal device 700 transmits the edited in-house information to the information processing device 200 (step S1815). When the information processing device 200 receives the edited in-house information, it transmits the edited in-house information to the second storage device 400 (step S1816).

[0153] The second storage device 400 causes the storage unit 440 to store the edited in-house information in the document storage unit 410 (step S1817). More specifically, the storage unit 440 may overwrite the edited in-house information in the document storage unit 410 with the in-house information selected as the subject of editing.

[0154] The following describes an editing screen displayed on the terminal device 700. Fig. 19 is a diagram showing an example of a display on the terminal device of the third embodiment.

[0155] A screen 191 shown in FIG. 19 is an example of an edit screen displayed on the terminal device 700 in step S1813 of FIG.

[0156] Screen 191 includes a display area 191a and operation buttons 191b. Display area 191a may display in-house information selected in step S1810 of Fig. 18. Furthermore, the in-house information displayed in display area 191a may be, for example, information obtained by converting an in-house document printed on paper into text data by OCR (Optical Character Recognition) processing or the like.

[0157] In this embodiment, the user of the terminal device 700 can edit the company information (text data) displayed in the display area 191a by manual input or the like.

[0158] Operation button 191b is an operation button for transmitting the text data displayed in display area 191a to information processing device 200 as edited in-house information.

[0159] In this embodiment, when the operation button 191b is operated, the terminal device 700 transmits the in-house information displayed in the display area 191a at the time the operation button 191b is operated to the information processing device 200 as edited in-house information.

[0160] In this embodiment, by doing so, for example, even if the OCR process fails to read the data or if there is a recognition error, the administrator can correct the error. Therefore, according to this embodiment, the in-house information that serves as the information source to be provided to the server device 500 can be made more accurate. Therefore, according to this embodiment, it is possible to prevent erroneous output response data from being output in response to question data.

[0161] (Fourth embodiment) The fourth embodiment will be described below with reference to the drawings. The fourth embodiment differs from the first embodiment in that it is possible to set information to be referenced preferentially in the information source provided to the server device 500. 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 description of the first embodiment, and their description will be omitted.

[0162] Fig. 20 is a sequence diagram illustrating the operation of the chat system of the fourth embodiment. The processing from step S2001 to step S2003 in Fig. 20 is the same as the processing from step S1801 to step S1803 in Fig. 18, so the description will be omitted.

[0163] It is assumed that in step S2003 in FIG. 20, an operation to add in-house information has been performed on the management screen.

[0164] The terminal device 700 accepts an operation to select the internal company information to be uploaded (step S2004), and when it accepts an instruction to set priority information that should be referenced preferentially for the selected internal company information (step S2005), it sends a request to the information processing device 200 to display a priority information setting screen (step S2006).

[0165] When the information processing device 200 receives the request to display the priority information setting screen, it transmits an instruction to display the priority information setting screen to the terminal device 700 (step S2007). The terminal device 700 receives the display instruction and displays the priority information setting screen (step S2008).

[0166] Next, the terminal device 700 accepts an operation to set priority information (step S2009), and when the setting is completed, transmits the company information and priority information to be added to the information processing device 200 (step S2010). At this time, the company information and priority information are assigned the tenant ID of the tenant to which the user of the terminal device 700 belongs.

[0167] When the information processing device 200 receives the in-house information and priority information to which the tenant ID is assigned, it transmits them to the second storage device 400 (step S2011).

[0168] When the second storage device 400 receives the in-house information and priority information to which the tenant ID has been assigned, the storage unit 440 stores the in-house information and priority information to which the tenant ID has been assigned in the document storage unit 410.

[0169] The priority information setting screen will be described below with reference to Fig. 21. Fig. 21 is a diagram showing an example of a display on a terminal device according to the fourth embodiment.

[0170] A screen 181 shown in FIG. 21 is an example of a setting screen for priority information displayed on the terminal device 700 in step S2009 of FIG.

[0171] Screen 181 includes input field 181a and operation buttons 181b. Input field 181a is an input field for inputting priority information, and operation button 181b is an operation button for storing the priority information input in input field 181a in association with in-house information.

[0172] In this embodiment, after the internal company information to be uploaded is selected, priority information is entered in the input field 181a on the screen 181, and when the operation button 181b is operated, the entered priority information is associated with the selected internal company information and transmitted to the information processing device 200.

[0173] Next, with reference to FIG. 22, a prompt when priority information is associated with in-house information will be described.

[0174] Fig. 22 is a diagram showing an example of a prompt in the fourth embodiment. The prompt 22 shown in Fig. 22 is an example of a prompt requesting generation of an answer, which is transmitted from the information processing device 200 to the server device 500 in step S910 of Fig. 9, for example.

[0175] The prompt 22 includes text data 22a, 22b, 22c, and 22d. The text data 22a indicates a request to generate a reply based on an information source in response to an inquiry from the user of the terminal device 600, and to refer to the priority information if priority information is associated with the information source.

[0176] The text data 22b indicates the content of an inquiry from the user of the terminal device 600. The text data 22c indicates an information source. The information source indicated by the text data 22c is the in-house information acquired from the second storage device 400 in step S909 of Fig. 9. The text data 22d indicates priority information associated with the information source indicated by the text data 22c.

[0177] In this embodiment, by inputting such a prompt 22 into the large-scale language model 511 of the server device 500, it is possible to obtain output response data that prioritizes preset priority information.

[0178] Therefore, according to this embodiment, it is possible to obtain output response data generated based on an appropriate information source for question data, and it is possible to prevent output of output response data with incorrect content.

[0179] (Fifth embodiment) The fifth embodiment will be described below with reference to the drawings. The fifth embodiment differs from the first embodiment in that the question and answer storage unit 320 is searched before requesting an answer from the server device 500. In the following description of the fifth 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 description of the first embodiment, and their description will be omitted.

[0180] 23 is a sequence diagram illustrating the operation of the chat system of the fifth embodiment. In the chat system 100 of this embodiment, when the information processing device 200 receives input of question data from the terminal device 600 (step S2301), the search request unit 252 transmits a search request to the first storage device 300 (step S2302).

[0181] The search request sent to the first storage device 300 includes a tenant ID and question data, and requests a search of the question and answer storage unit 320 in which question and answer information associated with the tenant ID is stored.

[0182] The first storage device 300 receives this search request, and causes the search unit 330 to search the question and answer storage unit 320 in which the question and answer information associated with the tenant ID is stored (step S2303), and transmits the registered answer data extracted as the search result to the information processing device 200 (step S2304). Specifically, the search unit 330 searches for the question and answer information associated with the tenant ID, and extracts, from the question and answer information, the registered answer data associated with the question indicated by the question data, as the search result.

[0183] Here, if the registered answer data is not extracted, that is, if the question and answer storage unit 320 does not store a question corresponding to the question data input in step S2301, the information processing device 200 proceeds to step S902 in Figure 9 and performs processing from step S902 onwards.

[0184] Furthermore, if one or more pieces of registered response data are extracted, the information processing device 200 performs the processes from step S2305 onwards, which will be described later.

[0185] The information processing device 200 transmits the registered answer data acquired from the first storage device 300 to the terminal device 600 (step S2305). The terminal device 600 causes the display control unit 650 to display the registered answer data acquired from the information processing device 200 as an answer to the question data (step S2306).

[0186] In this case, the registered answer data displayed on the terminal device 600 is extracted from the question and answer information stored in the question and answer storage unit 320.

[0187] In this embodiment, for example, when an instruction to obtain an answer using the large-scale language model 511 is given in the terminal device 600, the processes from step S2307 onwards may be performed.

[0188] The terminal device 600 receives, via the input receiving unit 640, an operation to instruct acquisition of a response from the server device 500 on the screen displayed in step S2306 (step S2307). Subsequently, the terminal device 600 transmits to the information processing device 200 a message indicating that acquisition of a response from the server device 500 has been instructed (step S2308).

[0189] When the information processing device 200 receives the instruction to acquire a response from the server device 500 in step S2308, the information processing device 200 performs the processes from step S902 onward in FIG.

[0190] Display examples of the terminal device 600 of this embodiment will be described below with reference to FIGS.

[0191] Fig. 24 is a first diagram showing a display example of a terminal device according to the fifth embodiment. Screen 240 shown in Fig. 24 is an example of a screen displayed on terminal device 600 when no registered response data is extracted in step S2304 of Fig. 23.

[0192] For example, in the chat system 100, the screen 240 may be displayed on the terminal device 600 in step S915 of FIG. 9 when the processing from step S902 onwards of FIG. 9 is executed after the processing up to step S2304 of FIG. 23 has been executed.

[0193] The screen 240 includes display areas 240a and 240b. The display area 240a displays question data input by the user of the terminal device 600. The display area 240b displays answer data output from the chat system 100.

[0194] 24, display area 240b displays text data indicating that no answer corresponding to the question indicated by the question data displayed in display area 240a is stored in question and answer storage unit 320. Display area 240b also displays text data indicating that the answer obtained from server device 500 having large-scale language model 511 for the question data is set as output answer data.

[0195] In this embodiment, even if the question and answer storage unit 320 does not store registered answer data corresponding to the question indicated by the question data, the output answer data generated by the server device 500 can be presented to the user.

[0196] Fig. 25 is a second diagram showing a display example of a terminal device according to the fifth embodiment. Screen 240A shown in Fig. 25 is an example of a screen displayed on terminal device 600 in step S2306 when one piece of registered response data is acquired in step S2304 of Fig. 23.

[0197] Screen 240 includes display areas 240a, 240c, and 240d. Display area 240c displays registered answer data included in the question and answer information stored in question and answer storage unit 320. In other words, display area 240c displays the registered answer data acquired as the search result in step S2304.

[0198] The display area 240d displays text data indicating that output answer data generated by the large-scale language model 511 possessed by the server device 500 can be obtained, and an operation button 240d-1 that instructs the user to obtain the answer generated by the large-scale language model 511 (generation AI) possessed by the server device 500.

[0199] In this embodiment, when the operation button 240d-1 displayed in the display area 240d is operated, the processes from step S2307 onward in FIG. 23 may be performed.

[0200] Fig. 26 is a third diagram showing a display example of a terminal device according to the sixth embodiment. Screen 240B shown in Fig. 26 is an example of a screen displayed on terminal device 600 in step S2306 when multiple pieces of registered response data are acquired in step S2304 of Fig. 23.

[0201] The screen 240B includes display areas 240a and 240e. In the display area 240e, a plurality of registered answer data 240e-1 included in the answer to question information stored in the question and answer storage unit 320, and an operation button 240d-1 are displayed.

[0202] In other words, the plurality of registered response data 240e-1 is the plurality of registered response data acquired as the search result in step S2304 in the display area 240e.

[0203] In this embodiment, when the operation button 240d-1 is operated on the screen 240B, the processes from step S2307 onward in FIG. 23 may be performed.

[0204] In this embodiment, by displaying the operation button 240d-1 in this manner, it is possible to cause the server device 500 to generate output response data in response to the user's request.

[0205] Furthermore, in this embodiment, before the server device 500 generates an answer, the question and answer storage unit 320 is searched and registered answer data associated with the registered question data corresponding to the question indicated by the question data is output. In this case, the registered answer data is an appropriate answer to the question indicated by the question data. In this embodiment, by combining the question and answer storage unit 320 and the server device 500 to output answers to questions, it is possible to reduce the frequency with which incorrect or inappropriate answers are output.

[0206] (Sixth embodiment) A sixth embodiment will be described below with reference to the drawings. The sixth embodiment differs from the first embodiment in that it is possible to switch whether or not to use information other than internal company information as an information source. In the following description of the sixth 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 description of the first embodiment, and descriptions thereof will be omitted.

[0207] FIG. 27 is a sequence diagram illustrating the operation of the chat system according to the sixth embodiment.

[0208] When the information processing device 200 receives input of question data from the terminal device 600 (step S2701), the prompt generating unit 255 refers to the setting information related to the information source to be given to the large-scale language model 511 (step S2702).

[0209] Specifically, the prompt generation unit 255 refers to setting information that indicates whether the information sources to be provided to the large-scale language model 511 are to be internal company information and information published on the Internet, or only information published on the Internet.

[0210] The setting information of this embodiment may be input by, for example, a user of the terminal device 600. In this case, the setting information may be input when the question data is input into the terminal device 600, and may be transmitted to the information processing device 200 together with the question data.

[0211] The setting information of this embodiment may be stored in advance in the information processing device 200 by a user of the terminal device 700 who is an administrator of in-house information. The setting information may also be a setting that sets only in-house information as an information source to be provided to the large-scale language model 511. In this case, the information processing device 200 may execute the processes from step S902 onwards in FIG. 9 after step S2702.

[0212] Furthermore, the process of referencing the setting information may be executed at a timing different from step S2702. Specifically, the setting information may be referenced before step S2701.

[0213] If the information processing device 200 is set to provide the information sources to the large-scale language model 511 with internal company information and information published on the Internet, the information processing device 200 proceeds to step S2703, and the chat system 100 performs processing from step S2703 to step S2714 to obtain output response data to be output to the terminal device 600.

[0214] The processing from step S2703 to step S2714 in Fig. 27 is the same as the processing from step S902 to step S913 in Fig. 9 except that the prompt generated in step S2712 is different from the prompt generated in step S911 in Fig. 9, and therefore description thereof will be omitted. Details of the prompt generated in step S2712 will be described later.

[0215] Furthermore, if the information processing device 200 is set to provide only information published on the Internet as information sources to the large-scale language model 511, the information processing device 200 proceeds to step S2715, and the chat system 100 performs processing from step S2715 to step S2717 and obtains output response data from the server device 500.

[0216] The processing from step S2715 to step S2717 in Fig. 27 is the same as the processing from step S911 to step S913 in Fig. 9, except that the prompt generated in step S2715 is different from the prompt generated in step S911 in Fig. 9, so a description thereof will be omitted. Details of the prompt generated in step S2715 will be described later.

[0217] When the information processing device 200 acquires the output response data from the server device 500, the process proceeds to step S2718, and the chat system 100 performs the processes from step S2718 to step S2721.

[0218] The processing from step S2718 to step S2721 is the same as the processing from step S914 to step S917 in FIG. 9, and therefore a description thereof will be omitted.

[0219] Next, with reference to FIGS. 28 to 30, a prompt transmitted from the information processing device 200 to the server device 500 in this embodiment will be described.

[0220] 28 is a first diagram showing an example of a prompt according to the sixth embodiment. The prompt 28 shown in FIG. 28 is an example of a prompt generated by the prompt generating unit 255 in step S2715 of FIG.

[0221] The prompt 28 includes text data 28a and text data 28b. The text data 28a is text data requesting the generation of output answer data to the question indicated by the question data. The text data 28b is text data indicating the question data input in step S2701.

[0222] In this embodiment, by inputting such a prompt 28 into the large-scale language model 511, output response data can be obtained using information on the Internet as the information source used by the large-scale language model 511 for learning.

[0223] 29 is a second diagram showing an example of a prompt according to the sixth embodiment. The prompt 29 shown in FIG. 29 is an example of a prompt generated by the prompt generating unit 255 in step S2712 of FIG.

[0224] Prompt 29 includes text data 29a, 28a, and 29b. Text data 29a is text data that requests the generation of output response data based on in-house information shown as text data 29b and information on the Internet that large-scale language model 511 used for training.

[0225] The text data 29b is the in-house information acquired by the information processing device 200 in step S2711 of FIG.

[0226] In this embodiment, by inputting such a prompt 29 into the large-scale language model 511, it is possible to present to the user an answer generated from in-house information and publicly available information.

[0227] 30 is a third diagram showing an example of a prompt according to the sixth embodiment. The prompt 30 shown in FIG. 30 is another example of the prompt generated by the prompt generating unit 255 in step S2712 of FIG.

[0228] Prompt 30 includes text data 30a, 28a, and 29b. Text data 30a is text data that requests the generation of output answer data using in-house information shown as text data 29b, and if the answer is not found in the in-house information, requests the generation of output answer data using information on the Internet that large-scale language model 511 used for training.

[0229] In this embodiment, by inputting such a prompt 30 into the large-scale language model 511, even if the answer is not contained in the internal company information used as the source of the information, an answer generated from publicly available information can be presented to the user.

[0230] In this manner, in this embodiment, it is possible to set whether the information source when generating output response data is to be internal company information and publicly available information, or to be only publicly available information. Therefore, in this embodiment, if a user of chat system 100 desires a general response, the information source when generating output response data may be only publicly available information. Also, if a user of chat system 100 desires a response from internal company information, the information source when generating output response data may be internal company information and publicly available information.

[0231] In this way, according to this embodiment, it is possible to present an answer based on an information source desired by the user.

[0232] (Seventh embodiment) The seventh embodiment will be described below with reference to the drawings. The seventh embodiment differs from the first embodiment in that the search target range is specified from within the company information in accordance with the user's operation or the content of the question. In the following description, the 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 description of the first embodiment, and their description will be omitted.

[0233] An overview of this embodiment will be described below. In this embodiment, an answer using internal company information desired by the user asking a question as an information source is generated in the large-scale language model 511. More specifically, in this embodiment, the user asking a question specifies the range of internal company information to be used as the information source for the answer, and the search results for the internal company information within the specified range are included in a prompt requesting the large-scale language model 511 to generate an answer.

[0234] Furthermore, in this embodiment, if the user asking the question does not specify the range of internal company information, the range of internal company information to be used as the source of the answer is specified according to the content of the question. Then, in this embodiment, the search results for the internal company information in the specified range are included in the prompt that requests the large-scale language model 511 to generate an answer.

[0235] In this embodiment, by doing so, it is possible to present the user who has asked the question with an answer that uses the company information desired by the user as the information source.

[0236] FIG. 31 is a diagram illustrating the functions of each device included in the chat system of the seventh embodiment.

[0237] First, the second storage device 400A included in the chat system 100A of this embodiment will be described.

[0238] The second storage device 400A of this embodiment includes a document storage unit 410A, a chunk storage unit 415, a user information storage unit 420, and a scenario storage unit 425. The document storage unit 410A of this embodiment is provided with multiple storage areas, and internal company information is stored in each storage area according to settings made by an administrator of the chat system 100A, etc. In the following description, each of the multiple storage areas provided in the document storage unit 410A is referred to as a folder. The chunk storage unit 415 stores chunks for each internal company information.

[0239] The scenario storage unit 425 stores information for specifying the search range of the internal company information stored in the second storage device 400A. In the following description, the information stored in the scenario storage unit 425 is referred to as scenario information. The scenario information of this embodiment may be set in advance by an administrator of the chat system 100A or the like and stored in the scenario storage unit 425.

[0240] The document storage unit 410A, chunk storage unit 415, and scenario storage unit 425 will be described in detail later.

[0241] Next, a description will be given of functions of the information processing device 200A. The information processing device 200A of this embodiment includes an input receiving unit 251, a search request unit 252, a conversation history acquisition unit 253, an information acquisition unit 254, a prompt generation unit 255, a prompt output unit 256, an information output unit 257, a division unit 258, a scenario setting unit 259, a scenario acquisition unit 260, a search range specification unit 261, and a question analysis unit 262.

[0242] The dividing unit 258 divides the in-house information stored in the document storage unit 410A into multiple chunks. The scenario setting unit 259 sets scenario information. In other words, the scenario setting unit 259 generates scenario information in response to operations on the terminal device 700 and stores it in the scenario storage unit 425 of the second storage device 400A.

[0243] The scenario acquisition unit 260 acquires scenario information stored in the scenario storage unit 425 of the second storage device 400A in response to the operation of the terminal device 600. The search range specification unit 261 specifies the search range of chunks stored in the chunk storage unit 415 based on the scenario information acquired by the scenario acquisition unit 260. The question analysis unit 262 analyzes question data input from the terminal device 600. Specifically, the question analysis unit 262 performs morphological analysis on the question data and acquires attributes of the user associated with the terminal device 600 to which the question data was input, and specifies the scenario information to be referenced.

[0244] Next, the document storage unit 410A, chunk storage unit 415, and scenario storage unit 425 of the second storage device 400A of this embodiment will be described with reference to FIGS.

[0245] 32 is a diagram showing an example of a document storage unit of the seventh embodiment. In the document storage unit 410A of this embodiment, a plurality of folders are created, and in-house information is stored in each folder. The folders may be created in advance by an administrator of the chat system 100A, etc.

[0246] In document storage unit 410A, the file name of the in-house information to be stored, the name of the folder in which the in-house information is stored, the update date and time of the in-house information, etc. may be stored in association with the in-house information.

[0247] In the example of Figure 32, it can be seen that the internal information with the file name "Sales Report" was stored in the "Sales Materials" folder at 9:00 on June 24, 2024. Also, in the example of Figure 32, it can be seen that the internal information with the file name "Regarding Paid Leave" was stored in the "Development Department" folder at 9:00 on June 18, 2024.

[0248] The folders created in document storage unit 410A may be created for each business operation of the tenant or each department of the tenant, for example. The folders in document storage unit 410A can be created arbitrarily by the administrator of the tenant who introduces chat system 100.

[0249] Fig. 33 shows an example of the chunk storage unit of the seventh embodiment. In Fig. 33, the file name of the in-house information as information for identifying the in-house information, the update date and time of the in-house information, the chunk ID for identifying the chunk, and the chunk are associated with each other.

[0250] In the example of FIG. 33, one page of internal information in the internal document "Sales Report" indicated by the internal information is treated as a chunk, and the number of pages is used as a chunk ID for identifying the chunk.

[0251] It should be noted that a chunk is not limited to one page of in-house information, but may be, for example, each paragraph or each chapter in an in-house document.

[0252] 34 is a diagram illustrating an example of a scenario storage unit according to the seventh embodiment. Scenario information according to this embodiment may be stored for each tenant.

[0253] In the scenario information stored in the scenario storage unit 425 of this embodiment, the tenant ID, the specified item, and the search target folder are associated as information items. The value of the item "specified item" indicates an item that allows the user to specify the range of internal company information to be used as the source of the answer. In other words, the value of the item "specified item" is an item for classifying internal company information, and the internal company information is classified by the specified item and stored in the document storage unit 410A.

[0254] In this embodiment, a list of items indicated by the item "Specified Item" is displayed on the terminal device 600 as options to be selected when specifying the range of in-house information. The value of the item "Search Target Folder" indicates the folder name associated with the specified item.

[0255] In the example of Fig. 34, the designated items include "Sales Department," "Development Department," and "Taking Paid Leave," which are displayed on the terminal device 600. Also, in the example of Fig. 34, the designated item "Sales Department" is associated with the "Sales Materials" folder, the designated item "Development Department" is associated with the "Development Department" folder, and the designated item "Taking Paid Leave" is associated with the "Attendance Management" folder and the "Labor Management" folder.

[0256] 34, operation buttons such as "Sales Department," "Development Department," and "Take Paid Leave" may be displayed on terminal device 600 as options to be selected when selecting a specified item. When the operation button indicating "Sales Department" is selected from the operation buttons displayed as options, the company information stored in the "Sales Department" folder in document storage unit 410A is specified as the information source of the response.

[0257] 34, the specified item is set separately from the folder name provided in document storage unit 410A, but is not limited to this. In this embodiment, the folder name provided in document storage unit 410A may also be used as the specified item.

[0258] In this case, terminal device 600 may display a list of folder names created in document storage unit 410A as a list of items for specifying the range of in-house information to be used as an information source.

[0259] 34, the scenario information is information that associates designated items with folders to be searched, but is not limited to this. The scenario information may include operation information that specifies the operation of the chat system 100A for each designated item. The operation information may be, for example, information indicating the operation of the chat system 100A when there is insufficient information to be obtained from a question input from the terminal device 600. The operation information may also be information indicating the operation of the chat system 100A when there are multiple folders associated with the designated items.

[0260] Next, the operation of the chat system 100A when storing in-house information in the second storage device 400A will be described with reference to Fig. 35. Fig. 35 is a first sequence diagram illustrating the operation of the chat system of the seventh embodiment.

[0261] In the chat system 100, the terminal device 700 causes the display control unit 750 to display an administrator screen for the administrator (step S3501). Next, the terminal device 700 causes the input receiving unit 740 to receive an operation to select a folder in which to store the internal company information (step S3502). Next, the terminal device 700 causes the input receiving unit 740 to receive a selection of the internal company information to be stored in the folder (step S3503). Next, the terminal device 700 causes the communication control unit 730 to transmit (upload) the selected internal company information to the information processing device 200 (step S3504).

[0262] When information processing device 200A receives the in-house information via input receiving unit 251, information output unit 257 transmits the received in-house information to second storage device 400A (step S3505), and transmits a reception notification indicating that the in-house information has been received to terminal device 700 (step S3506). Second storage device 400A stores the in-house information via storage unit 440 in the selected folder in document storage unit 410A (step S3507).

[0263] Next, the information processing device 200A divides the in-house information into chunks using the dividing unit 258 (step S3508). Next, the information processing device 200A transmits the divided chunks to the second storage device 400A using the information output unit 257 (step S3509).

[0264] The second storage device 400A stores the chunk received from the information processing device 200A in the chunk storage unit 415 by the storage unit 440 together with information specifying the in-house information containing the chunk (step S3510).

[0265] Next, the information processing device 200A causes the information output unit 257 to transmit a completion notification indicating that the registration of the in-house information has been completed to the terminal device 700 (step S3511).

[0266] Next, the operation of the chat system 100A when storing scenario information in the scenario storage unit 425 will be described with reference to Fig. 36. Fig. 36 is a second sequence diagram illustrating the operation of the chat system of the seventh embodiment.

[0267] In the chat system 100A of this embodiment, the terminal device 700 causes the display control unit 750 to display an administrator screen for the administrator (step S3601). Next, the terminal device 700 causes the input receiving unit 740 to receive an operation for setting scenario information (step S3602). Next, the terminal device 700 causes the communication control unit 730 to transmit a request to display a scenario setting screen to the information processing device 200A (step S3603).

[0268] When the information processing device 200A receives this display request, the scenario setting unit 259 outputs a display instruction for the scenario setting screen to the terminal device 700 via the information output unit 257 (step S3604).

[0269] The terminal device 700 receives this display instruction and causes the display control unit 750 to display the scenario setting screen (step S3605). Subsequently, the terminal device 700 receives input of scenario information through the input receiving unit 740 (step S3606).

[0270] For example, when a specified item is entered on a scenario setting screen and a folder name to be associated with the specified item is selected, the terminal device 700 may associate the specified item, the selected folder name, and the tenant ID of the tenant to which the user of the terminal device 700 belongs, and use this as scenario information.

[0271] The specified items input here may be input by the user of the new terminal device 700 or may be selected from the scenario information stored in the scenario storage unit 425.

[0272] Next, the terminal device 700 transmits the input scenario information to the information processing device 200A via the communication control unit 730 (step S3607).

[0273] When the scenario setting unit 259 of the information processing device 200A receives the scenario information from the terminal device 700, the scenario setting unit 259 outputs the received scenario information to the second storage device 400A via the information output unit 257 (step S3608).

[0274] The second storage device 400A stores the scenario information received from the information processing device 200A in the scenario storage unit 425 (step S3609) and notifies the information processing device 200A of the completion of storage (step S3610). Upon receiving this notification, the information processing device 200A transmits, via the information output unit 257, a notification indicating that reception of the scenario information has been completed to the terminal device 700 (step S3611). The terminal device 700 causes the display control unit 750 to display the notification received from the information processing device 200A (step S3612).

[0275] In this embodiment, by setting the scenario information in this manner through the above processing, a list of designated items to be displayed on the terminal device 600 can be created.

[0276] Next, a display example of the terminal device 700 will be described with reference to Fig. 37. Fig. 37 is a first diagram showing a display example of the terminal device in the seventh embodiment. Screen 140 shown in Fig. 37 is an example of an administrator screen displayed on the terminal device 700 in step S3501 of Fig. 35 and step S3601 of Fig. 36.

[0277] Screen 140 includes a display area 141 and operation buttons 142. Display area 141 displays a list of in-house information stored in a folder provided in document storage unit 410A. In the example of Fig. 14, a list of in-house information stored in the "Sales Department" folder is displayed.

[0278] The operation button 142 is an operation button for displaying a scenario setting screen. In this embodiment, when the operation button 142 is operated on the screen 140, a request to display the scenario setting screen is transmitted from the terminal device 700 to the information processing device 200A, and the screen 140 transitions to the scenario setting screen.

[0279] Fig. 38 is a second diagram showing an example of a display on a terminal device in the seventh embodiment. Screen 150 shown in Fig. 38 is an example of a scenario setting screen displayed on terminal device 700 in step S3604 of Fig. 36. Screen 150 is also an example of a scenario setting screen for setting scenario information for the specified item "Sales Department."

[0280] Screen 150 also includes display areas 151 and 152, and operation buttons 153. Display area 151 displays display area 154 for selecting an object to be associated with a specified item, and display area 154 includes operation buttons 154a and 154b.

[0281] The target associated with the specified item may be a folder to be searched, or may be operation information associated with each specified item. Operation button 154a is an operation button for selecting a folder to be searched as a target to be associated with the specified item. Operation button 154b is an operation button for adding a new target to be associated with the specified item.

[0282] 38, operation button 154a is selected in display area 154, and a selection field for a folder to be associated with the designated item "Sales Department" is displayed in display area 152. Specifically, a "Sales Materials" folder is associated with the designated item "Sales Department" in display area 152. Operation button 153 is an operation button for storing scenario information including the designated item and folder name associated in display area 152 in second storage device 400A.

[0283] In this embodiment, when the operation button 153 is operated after the designated items are associated with the search target folder and the operation information on the screen 150, scenario information is transmitted from the terminal device 700 to the information processing device 200A.

[0284] In this embodiment, the administrator of chat system 100A can set the operation of chat system 100A when chat system 100A starts providing services.

[0285] Fig. 39 is a third diagram showing a display example of a terminal device according to the seventh embodiment. Screen 160 shown in Fig. 39 is an example of a screen for setting the operation of chat system 100A when chat system 100A starts providing a service.

[0286] The screen 160 includes setting fields 161 and 162. The setting field 161 is used to set a message to be displayed on the terminal device 600 when the chat system 100A starts providing a service. In the example of Fig. 39, it can be seen that the setting has been set to display the message "Do you have any questions? Enter your question directly or select a method of inquiry using the buttons."

[0287] In the setting field 162, it can be seen that when a specified item is selected in the list of specified items displayed on the terminal device 600 at the start of service provision by the chat system 100A, processing is performed in accordance with the scenario information associated with the specified item. In this embodiment, the administrator of chat system 100A can set the operation of chat system 100A when chat system 100A starts providing a service.

[0288] Next, the operation of the chat system 100A when a question is asked by the user of the terminal device 600 will be described with reference to Fig. 40. Fig. 40 is a third sequence diagram illustrating the operation of the chat system of the seventh embodiment.

[0289] In the chat system 100A, the terminal device 600 displays a chat screen by the display control unit 650 (step S4001). On this chat screen, a list of items designated by the user of the terminal device 600 is displayed as a list of operation buttons.

[0290] The following processing from step S4002 to step S4012 indicates the operation of the chat system 100A when an operation button is selected on the chat screen displayed on the terminal device 600 in step S4001.

[0291] When the terminal device 600 accepts the selection of an operation button displayed on the chat screen (step S4002), the communication control unit 630 transmits information identifying the selected operation button to the information processing device 200A (step S4003). In other words, when the terminal device 600 accepts the selection of a designated item from the list of designated items, it notifies the information processing device 200A of the selected designated item.

[0292] When the information processing device 200A receives information identifying the selected operation button (the selected specified item), the scenario acquisition unit 260 sends an acquisition request for scenario information corresponding to the specified item to the second storage device 400A (step S4004). When the second storage device 400A receives this acquisition request, it searches the scenario storage unit 425 for the specified item and acquires the scenario information corresponding to the specified item (step S4005). Subsequently, the second storage device 400A transmits the acquired scenario information to the information processing device 200A (step S4006).

[0293] When the information processing device 200A acquires the scenario information, the search range specification unit 261 specifies the in-house information in the search target folder included in the scenario information as the search range (step S4007). Subsequently, the information processing device 200A outputs, to the terminal device 600, a message indicating that a response based on the in-house information within the search range will be output (step S4008) through the information output unit 257.

[0294] When the terminal device 600 receives the input of the question data on the chat screen, the communication control unit 630 transmits the question data to the information processing device 200A (step S4009).

[0295] When the information processing device 200A receives input of question data via the input receiving unit 251, the search request unit 252 sends a search request within the search range identified in step S4007 to the second storage device 400A (step S4010). The search request sent to the second storage device 400A here may be the result of morphological analysis, etc., performed by the question analysis unit 262 on the question data input in step S4009.

[0296] When the second storage device 400A receives the search request, the search unit 430 searches for chunks stored in the chunk storage unit 415 that fall within the search range identified in step S4007 (step S4011), sends the search results to the information processing device 200A (step S4012), and proceeds to step S4022, which will be described later.

[0297] The following processing from step S4013 to step S4018 shows the operation of the chat system 100A when a question is directly input without selecting an operation button on the chat screen displayed on the terminal device 600 in step S4001.

[0298] When the question data is input, the terminal device 600 transmits the question data to the information processing device 200A via the communication control unit 630 (step S4013). When the information processing device 200A receives the input of the question data via the input receiving unit 251, the question analysis unit 262 analyzes the question data and identifies the specified items (step S4014). In other words, the question analysis unit 262 analyzes the question data and identifies the scenario information to be referenced based on the analysis results.

[0299] The following describes the processing of the question analysis unit 262. The question analysis unit 262 of this embodiment may, for example, refer to the scenario storage unit 425 of the second storage device 400A and identify the specified item based on the content of the question indicated by the question data.

[0300] Furthermore, the question analysis unit 262 may refer to the user information of the user of the terminal device 600 to acquire the user attributes, and identify the designated items based on the user attributes. The user attributes may indicate, for example, the department to which the user belongs. In this case, it is assumed that the information processing device 200A has previously acquired the user ID of the user of the terminal device 600 and the tenant ID of the tenant to which the user belongs. The question analysis unit 262 may refer to the scenario storage unit 425 of the second storage device 400A to identify the designated items based on the attributes of the user of the terminal device 600.

[0301] Next, the information processing device 200A causes the scenario acquisition unit 260 to transmit an acquisition request for scenario information of the identified specified item to the second storage device 400A (step S4015). The processing from step S4015 to step S4018 in Fig. 40 is similar to the processing from step S4004 to step S4007, and therefore a description thereof will be omitted. Furthermore, the processing from step S4019 to step S4021 is similar to the processing from step S4010 to step S4012 in Fig. 40, and therefore a description thereof will be omitted. Note that the search request transmitted to the second storage device 400A in step S4019 may be the result of morphological analysis, etc., performed by the question analysis unit 262 on the question data input in step S4013.

[0302] In this embodiment, by doing this, even if the user of the terminal device 600 does not select a specified item, the specified item can be automatically identified from the context of the question data and the user's attributes, thereby increasing the probability that answer data based on the information source desired by the user will be output.

[0303] Information processing device 200A uses prompt generating unit 255 of second storage device 400 to generate a prompt requesting generation of an answer to the question using the search results as an information source (step S4022).

[0304] More specifically, in step S4002, when the prompt generation unit 255 receives input of a specified item indicating the classification of internal company information, it sets the chunk obtained as the search result of internal company information classified into the input specified item as internal company information to be included in a prompt requesting the generation of an answer to a question.

[0305] Furthermore, in step S4013, when question data is input, the prompt generation unit 255 uses the chunks obtained as the search results for the internal company information classified into the specified items identified by the question analysis unit 262 as internal company information to be included in a prompt requesting the generation of an answer to the question.

[0306] The processing from step S4022 to step S4028 in FIG. 40 is similar to the processing from step S911 to step S917 in FIG. 9, and therefore description thereof will be omitted.

[0307] The chat screen of this embodiment will be described below with reference to Fig. 41. Fig. 41 is a fourth diagram showing a display example of a terminal device of the seventh embodiment. Screen 170 shown in Fig. 41 is an example of a chat screen displayed on terminal device 600 in step S4026 of Fig. 41. Screen 170 shown in Fig. 41 is also an example of a chat screen displayed in step S4026 when a specified item is selected from the list of specified items in step S4001 of Fig. 40.

[0308] Screen 170 includes display areas 171, 172, 173, 174, and 175. Display area 171 includes display areas 171a and 171b. Display area 171a displays a message prompting the user to select a specified item or input question data. Note that the message displayed here may be set in advance as part of the operation information included in the scenario information.

[0309] Display area 171b displays a list of designated items and corresponding operation buttons. In other words, display area 171b displays a plurality of operation buttons as options for selecting designated items.

[0310] The designated item selected in the display area 171b is displayed in the display area 172. In the example of Fig. 41, it can be seen that the designated item "Development Department" has been selected.

[0311] The display area 173 includes a display area 173a and an operation button 173b. The display area 173a displays a message transmitted to the terminal device 600 in step S4008 of FIG. 40. More specifically, the display area 173a displays a message indicating an answer to a question regarding the specified item "Development Department." The operation button 173b is an operation button for redoing the question from the beginning. When the operation button 173b is selected, the display from the display area 172 onwards may be hidden.

[0312] Display area 174 displays question data entered by the user. Display area 175 includes display areas 175a, 175b, and 175c and operation button 175d. Display area 175a displays answer data output from large-scale language model 511. Display area 175b displays information indicating the storage location of the in-house information that served as the information source.

[0313] Display area 175c displays a message indicating that answer data has been generated by large-scale language model 511. Operation button 175d is an operation button for the user to end a conversation using chat system 100A.

[0314] The display area 171 on the screen 170 in FIG. 41 may be displayed on the terminal device 600, for example, when question data is directly input by the user in step S4001 in FIG. 40 and no specified item is identified by the processing of the question analysis unit 262 in step S4014.

[0315] Furthermore, in this embodiment, when question data is directly input by the user and a designated item is not identified by the processing of the question analysis unit 262 in step S4014, the information processing device 200A may cause the terminal device 600 to display a message asking about the attributes of the user. Note that, when a designated item is not identified by the processing of the question analysis unit 262, the information processing device 200A may cause the terminal device 600 to display a message including a question for prompting the user to identify a designated item.

[0316] Furthermore, when question data is directly input by the user and the specified item is not identified by the processing of the question analysis unit 262 in step S4014, the information processing device 200A may cause the terminal device 600 to display a question sentence for obtaining an additional question from the user. In this case, the information processing device 200A may generate a prompt requesting the generation of a question sentence for prompting the user of the terminal device 600 to input further information, output the prompt to the large-scale language model 511, and transmit the question sentence output from the large-scale language model 511 to the terminal device 600.

[0317] In this embodiment, the internal company information is classified in advance by designated item and stored in the document storage unit 410A. Then, in this embodiment, by having the user of the terminal device 600 select a designated item, an answer to the question can be generated by searching the internal company information within the range desired by the user. Therefore, according to this embodiment, it is possible to increase the likelihood that an appropriate answer will be output in response to the user's question, and to improve the accuracy of the answer to the question.

[0318] Furthermore, in this embodiment, a specified item indicating a classification of internal company information is identified based on the results of explaining the question data input from the terminal device 600, and the internal company information identified and associated with the specified item is used as a search target when acquiring information sources to be used in generating an answer. Therefore, according to this embodiment, even if the user does not select a specified item, an answer to a question can be generated by searching internal company information within the range desired by the user, thereby improving the accuracy of the answer to the question.

[0319] In addition, in the above-described embodiments, an example of user-specific information has been described using in-house information, but user-specific information is not limited to this. User-specific information can also be applied to, for example, store information or user-specific information. Store information includes store identification information (store name, address, telephone number), purchase history information, etc.

[0320] Furthermore, the user-specific information includes personal identification information, account information, purchase history information, browsing history information, location information, preference information, activity information, and the like. Personally identifiable information includes information that identifies a specific individual, such as name, address, telephone number, or email address. Account information includes information related to a specific online account, such as username, password, or account number. Purchase history information includes information such as the products or services a user has purchased, the date and time of purchase, and payment method. Browsing history information includes information such as the pages a user has viewed on websites or within apps, search history, and click history. Location information includes location-related information such as a user's current location and past movement history. Preference information is information that indicates content preferences and tastes, and includes information about content, products, and services that may be of interest to a user. Activity information is information about social media activity, such as the posts a user has shared, likes, and comments. This includes information such as your account history and who you follow.

[0321] 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.

[0322] Additionally, the devices described in each embodiment represent only one of multiple computing environments for implementing the embodiments disclosed herein.

[0323] 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.

[0324] 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.

[0325] 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.

[0326] For example, aspects of the present invention are as follows. <1> an input receiving unit that receives input of question data; a prompt output unit that, upon receiving the input of the question data, outputs a prompt to a large-scale language model requesting the generation of answer data to the question data based on information unique to the user; an information output unit that outputs answer data output from the large-scale language model based on the prompt output from the prompt output unit to a terminal device as output answer data. <2> a conversation history acquisition unit that acquires, from a first storage device, conversation history information that continued immediately before the question data was input; The prompt output unit outputting a prompt to the large-scale language model requesting the generation of a query for searching a second storage device storing the user-specific information, the query including the conversation history information acquired by the conversation history acquisition unit; <1> The information processing device described. <3> an information acquisition unit that acquires user-specific information that is a search result of the second storage device based on a query output from the large-scale language model in response to an input of a prompt requesting the generation of the query; The prompt output unit the user-specific information to be included in the prompt requesting the generation of the answer data is the user-specific information acquired by the information acquisition unit. <2> The information processing device described. <4> The prompt output unit When the conversation history information is not acquired by the conversation history acquisition unit, user-specific information acquired as a result of searching the second storage device using the question data is set as the user-specific information to be included in the prompt requesting generation of the answer data. <2> The information processing device described. <5> The prompt output unit outputting a prompt to the large-scale language model, the prompt including answer data output from the large-scale language model, and requesting correction of the answer data; <1> ~ <4> 10. The information processing device according to claim 9, wherein: <6> an information output unit that receives input of priority information associated with the user-specific information and stores the user-specific information in a storage device in association with the priority information; The prompt output unit If the priority information is associated with the user-specific information, a prompt is output to the large-scale language model requesting generation of the answer data that includes the priority information and prioritizes the priority information. <1> ~ <5> 10. The information processing device according to claim 9, wherein: <7> a search request unit that transmits a search request based on the question data to a first storage device that stores question and answer information in which the registered question data and the registered answer data are associated with each other; The prompt output unit if the first storage device does not store registered answer data corresponding to the question data, outputting a prompt to the large-scale language model requesting generation of the answer data; <1> ~ <6> 10. The information processing device according to claim 9, wherein: <8> The prompt output unit the first storage device stores registered answer data corresponding to the question data, and when an instruction to generate an answer using the large-scale language model is received, a prompt requesting generation of the answer data is output to the large-scale language model; <7> The information processing device described. <9> The prompt output unit outputting a prompt to the large-scale language model requesting generation of the answer data based on the user-specific information and information learned by the large-scale language model; <1> ~ <8> 10. The information processing device according to claim 9, wherein: <10> The prompt output unit If the user-specific information does not include answer data corresponding to the question data, a prompt is output to the large-scale language model to request generation of the answer data based on information learned by the large-scale language model. <1> ~ <8> 10. The information processing device according to claim 9, wherein: <11> a prompt generating unit that, when receiving input of the question data, generates a prompt requesting generation of answer data for the question data; <1> ~ <10> 10. The information processing device according to claim 9, wherein: <12> the user-specific information is stored in a second storage device, classified by designated item; The input receiving unit Accepting input of items indicating classification of the user's unique information; The prompt output unit a search result obtained by searching for the user-specific information classified into the items using the question data is used as the user-specific information to be included in a prompt requesting generation of the answer data; <1> The information processing device described. <13> the user-specific information is stored in a second storage device, classified by designated item; a question analysis unit that identifies the item corresponding to the question data based on a result of analyzing the question data; The prompt output unit a search result obtained by searching for the user-specific information classified into the category identified by the question analysis unit using the question data is used as the user-specific information to be included in a prompt requesting generation of the answer data; <1> The information processing device described. <14> 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 question data from the terminal device; a prompt output unit that, upon receiving the input of the question data, outputs a prompt to a large-scale language model requesting the generation of answer data to the question data based on information unique to the user; an information output unit that outputs answer data output from the large-scale language model based on the prompt output from the prompt output unit to the terminal device as output answer data, The terminal device an information processing system having a display control unit that displays the question data and the output response data in association with each other; <15> An information processing method including an information processing device and a terminal device that communicates with the information processing device, The information processing device, Accepting input of question data from the terminal device; outputting a prompt to the large-scale language model requesting generation of answer data for the question data based on user-specific information; outputting response data output from the large-scale language model based on the output prompt to the terminal device as output response data; The terminal device, The information processing method displays the question data and the output response data in association with each other. <16> Accepts input of question data, outputting a prompt to the large-scale language model requesting generation of answer data for the question data based on user-specific information; a program that causes an information processing device to execute a process of outputting answer data output from the large-scale language model based on the output prompt to a terminal device as output answer data;

[0327] 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]

[0328] 100 Chat System 200 Information processing device 251 Input Reception Unit 252 Search request section 253 Conversation History Acquisition Unit 254 Information Acquisition Department 255 Prompt Generation Unit 256 prompt output section 257 Information Output Unit 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]

[0329] [Patent Document 1] Japanese Patent Application Publication No. 2023-129333

Claims

1. an input receiving unit that receives input of question data; a prompt output unit that, upon receiving the input of the question data, outputs a prompt to a large-scale language model requesting the generation of answer data to the question data based on information unique to the user; an information output unit that outputs answer data output from the large-scale language model based on the prompt output from the prompt output unit to a terminal device as output answer data.

2. a conversation history acquisition unit that acquires, from a first storage device, conversation history information that continued immediately before the question data was input; The prompt output unit 2. The information processing device according to claim 1, wherein the information processing device outputs a prompt to the large-scale language model requesting the generation of a query for searching a second storage device that includes the conversation history information acquired by the conversation history acquisition unit and stores the user-specific information.

3. an information acquisition unit that acquires user-specific information that is a search result of the second storage device based on a query output from the large-scale language model in response to an input of a prompt requesting the generation of the query; The prompt output unit 3. The information processing apparatus according to claim 2, wherein the user-specific information included in the prompt requesting the generation of answer data is user-specific information acquired by the information acquisition unit.

4. The prompt output unit 3. The information processing device of claim 2, wherein when the conversation history information is not acquired by the conversation history acquisition unit, user-specific information acquired as a result of searching the second storage device using the question data is used as the user-specific information to be included in a prompt requesting the generation of the answer data.

5. The prompt output unit 3. The information processing apparatus according to claim 1, further comprising: outputting, to the large-scale language model, a prompt including answer data output from the large-scale language model and requesting correction of the answer data.

6. an information output unit that receives input of priority information associated with the user-specific information and stores the user-specific information in a storage device in association with the priority information; The prompt output unit 3. The information processing device according to claim 1, wherein, when the priority information is associated with the user-specific information, a prompt is output to the large-scale language model requesting the generation of the answer data that includes the priority information and prioritizes the priority information.

7. a search request unit that transmits a search request based on the question data to a first storage device that stores question and answer information in which the registered question data and the registered answer data are associated with each other; The prompt output unit 3. The information processing apparatus according to claim 1, wherein, when the first storage device does not store registered answer data corresponding to the question data, a prompt requesting generation of the answer data is output to the large-scale language model.

8. The prompt output unit 8. The information processing device according to claim 7, wherein the first storage device stores registered answer data corresponding to the question data, and when an instruction is given to generate an answer using the large-scale language model, a prompt requesting generation of the answer data is output to the large-scale language model.

9. The prompt output unit 2. The information processing apparatus according to claim 1, further comprising: a prompt for requesting generation of the response data, based on the user-specific information and information learned by the large-scale language model, outputting the prompt to the large-scale language model.

10. The prompt output unit 2. The information processing device according to claim 1, wherein, when the user-specific information does not include answer data corresponding to the question data, a prompt is output to the large-scale language model requesting generation of the answer data based on information learned by the large-scale language model.

11. 3. The information processing apparatus according to claim 1, further comprising a prompt generating unit configured to generate a prompt requesting generation of answer data for the question data when the question data is input.

12. the user-specific information is stored in a second storage device, classified by designated item; The input receiving unit Accepting input of items indicating classification of the user's unique information; The prompt output unit 2. The information processing device according to claim 1, wherein a search result obtained by searching for the user-specific information classified into the items using the question data is used as the user-specific information to be included in a prompt requesting generation of the answer data.

13. the user-specific information is stored in a second storage device, classified by designated item; a question analysis unit that identifies the item corresponding to the question data based on a result of analyzing the question data; The prompt output unit The information processing device according to claim 1, wherein the search results obtained by searching for the user-specific information classified into the items identified by the question analysis unit using the question data are used as the user-specific information to be included in a prompt requesting the generation of the answer data.

14. 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 question data from the terminal device; a prompt output unit that, upon receiving the input of the question data, outputs a prompt to a large-scale language model requesting the generation of answer data to the question data based on information unique to the user; an information output unit that outputs answer data output from the large-scale language model based on the prompt output from the prompt output unit to the terminal device as output answer data, The terminal device an information processing system having a display control unit that displays the question data and the output response data in association with each other;

15. An information processing method including an information processing device and a terminal device that communicates with the information processing device, The information processing device, Accepting input of question data from the terminal device; outputting a prompt to the large-scale language model requesting generation of answer data for the question data based on user-specific information; outputting response data output from the large-scale language model based on the output prompt to the terminal device as output response data; The terminal device, The information processing method displays the question data and the output response data in association with each other.

16. Accepts input of question data, outputting a prompt to the large-scale language model requesting generation of answer data for the question data based on user-specific information; a program that causes an information processing device to execute a process of outputting answer data output from the large-scale language model based on the output prompt to a terminal device as output answer data;

Images