Information processing device and information providing method

Abstract

To provide an information processing device and an information providing method capable of improving appropriateness of an answer to a question from a user.SOLUTION: An information processing device includes a storage unit in which a large scale language model is stored, and a processing circuit. The processing circuit performs verification on the basis of user information transmitted from a user terminal, accepts contents of a question input on the basis of an operation of the user terminal, and generates an answer to the contents of the question on the basis of a retrieval result from a prescribed database. A plurality of databases includes an occupied database to which an attribute label is imparted. The user information contains a user identification label associated with a user using the user terminal. The processing circuit determines the occupied database to which the attribute label associated with the user identification label contained in the user information in the plurality of databases is imparted as a prescribed database.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device and an information providing method. [Background technology]

[0002] A search system using a learning model has been proposed to provide information support to users who have installed a product in response to defects that have occurred in the product. For example, Patent Document 1 listed below describes a system in which past defect information and the like are compiled into a database, and a learning model is used to search the database based on the content of a question entered by a user to generate an answer. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2023-3933 Summary of the Invention [Problem to be solved by the invention]

[0004] However, in conventional search systems, the database to be searched is one, i.e., it is shared among multiple users. The product user who is asking the question may independently accumulate defect information, but such user-specific data may be confidential information that the user does not want other users to know. Therefore, the user may not want their own data to be included in the database to be searched in response to questions from other users. Therefore, conventional search systems have room for improvement in terms of improving the accuracy of answers without allowing others to use confidential information.

[0005] Such problems can occur not only in the search system for dealing with the above-mentioned product defects, but also in knowledge search systems in a wide range of fields, such as search systems for dealing with various business problems.

[0006] An object of the present disclosure is to provide an information processing device and an information providing method that can improve the accuracy of answers to user questions. [Means for solving the problem]

[0007] An information processing device according to one aspect of the present disclosure is an information processing device that can access multiple databases and communicate with multiple user terminals, and is equipped with a memory device in which a large-scale language model is stored, and a processing circuit, wherein the processing circuit performs authentication based on user information sent from the user terminal, accepts a question input based on operation of the user terminal, generates a search query for a specified database among the multiple databases using the large-scale language model from the question content, sends the search query to the specified database, generates an answer to the question from search results in the specified database using the large-scale language model, and outputs the answer to the user terminal, wherein the multiple databases include a proprietary database to which an attribute label is assigned, and the user information includes a user identification label associated with the user using the user terminal, and the processing circuit determines the proprietary database among the multiple databases to which the attribute label corresponding to the user identification label included in the user information is assigned as the specified database to which the search query is sent.

[0008] Another aspect of the present disclosure is an information provision method that generates an answer to a question sent from a user terminal and provides the answer, performing authentication based on user information sent from the user terminal, accepting the question input based on operation of the user terminal, generating a search query to a predetermined database from a plurality of databases using a large-scale language model from the question, sending the search query to the predetermined database, generating an answer to the question from search results in the predetermined database using the large-scale language model, and outputting the answer to the user terminal, wherein the plurality of databases include a proprietary database to which an attribute label is assigned, the user information includes a user identification label associated with a user using the user terminal, and determining the proprietary database from the plurality of databases to which the attribute label corresponding to the user identification label included in the user information is assigned as the predetermined database to which the search query is sent. [Effects of the Invention]

[0009] According to the present disclosure, it is possible to improve the accuracy of answers to user questions. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of a search system including an information processing device according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a block diagram showing a functional configuration of the information processing device shown in FIG. [Figure 3] FIG. 3 is a flowchart showing an example of the flow of information providing processing in the information processing device of this embodiment. [Figure 4] FIG. 4 is a block diagram showing a functional configuration of the information processing device shown in FIG. 1 when accessed from a user terminal different from that shown in FIG. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. Note that, in the following, the same or corresponding elements will be denoted by the same reference numerals throughout the drawings, and redundant description thereof will be omitted.

[0012] 1 is a block diagram showing a schematic configuration of a search system including an information processing device according to an embodiment of the present disclosure. The search system 1 in this embodiment includes an information processing device 2, a database group 3 communicatively connected to the information processing device 2, and a plurality of user terminals 4a, 4b communicatively connected to the information processing device 2. The information processing device 2 and the plurality of user terminals 4a, 4b are communicatively connected via a predetermined network N such as the Internet. The plurality of user terminals 4a, 4b are configured by computers capable of communicating with the information processing device 2, such as personal computers or smartphones.

[0013] The information processing device 2 includes a processing circuit 20 that performs various types of signal processing. The processing circuit 20 includes a computer such as a microcontroller, a personal computer, or a PLC (Programmable Logic Controller). More specifically, the processing circuit 20 includes a processor, a memory, and peripheral circuits. The processor includes, for example, a CPU or an MPU. The memory includes, for example, a ROM, a RAM, a register, and the like. The peripheral circuits include, for example, an input / output interface.

[0014] It should be noted that the functions of the elements disclosed herein can be performed using circuits or processing circuits, including general-purpose processors, special-purpose processors, integrated circuits, application-specific integrated circuits (ASICs), conventional circuits, and / or combinations thereof, configured or programmed to perform the disclosed functions. A processor is considered a processing circuit or circuit because it includes transistors and other circuitry. In this specification, a circuit, unit, means, or module is hardware that performs the recited functions or hardware that is programmed to perform the recited functions. The hardware may be hardware disclosed herein or other known hardware that is programmed or configured to perform the recited functions. Where hardware is a processor, which is considered a type of circuit, the circuit, unit, or means is a combination of hardware and software, and the software is used to configure the hardware and / or processor.

[0015] The information processing device 2 includes a memory 21 that stores a large-scale language model. The processing circuit 20 has a function called RAG (Retrieval Augmented Generation) and links the large-scale language model with a database group 3. The processing circuit 20 uses the large-scale language model to generate a search query for a predetermined database in response to a question entered by a user, and generates an answer to the question from the search results.

[0016] In this embodiment, the search system 1 is configured as a search system for dealing with defects in a predetermined product such as a robot. That is, the information processing device 2 receives a question about the defect in the product from the user terminal 4a, 4b, performs a search in the database group 3 using a large-scale language model, and outputs a solution to the defect as an answer to the question, thereby providing information to the user.

[0017] For this reason, the information processing device 2 is configured to be able to access the database group 3. The database group 3 includes multiple databases. The multiple databases include a common database 31 and proprietary databases 32a and 32b. The common database 31 includes a database related to product manuals or a database related to product defect information held by the product provider. The product provider may include not only the product manufacturer but also those who manage or maintain the product. The common database 31 may include a database related to product catalogs, specifications, etc. The product defect information in the common database 31 is generalized and does not include information specific to the product user. The common database 31 is a database managed by the product provider or the provider of the search system 1.

[0018] The proprietary databases 32a and 32b include databases related to product defect information held by product users, or databases related to the product usage status of the product users. The defect information contained in the proprietary databases 32a and 32b includes information held by the product users, such as the specific product state and usage status at the time of the defect. Furthermore, the product usage status contained in the proprietary databases 32a and 32b includes information on adjustable parameter values ​​of the product, the product usage environment, and the system configuration that includes the product as a part. For example, if the product is a robot, the product usage status may also include the robot's operating program, etc. The proprietary databases 32a and 32b are databases managed by the product users. In this way, the proprietary databases 32a and 32b may include confidential information held by the product users.

[0019] The proprietary databases 32a and 32b include a plurality of proprietary databases that differ for each user of the product. The plurality of proprietary databases 32a and 32b are assigned attribute labels, which are information that identifies the user of the product. In the example of FIG. 1, the proprietary database 32a managed by company A, which is a user of the product, is assigned attribute label LA, and the proprietary database 32b managed by company B, which is also a user of the product, is assigned attribute label LB. The common database 31 is not assigned an attribute label that identifies the user. However, the common database 31 may also be assigned an attribute label that indicates that it is a common database 31.

[0020] A user using user terminal 4a belongs to company A, which is a user of the product, and a user using user terminal 4b belongs to company B, which is also a user of the product. Therefore, the user information referenced when a user using user terminal 4a logs in to the information processing device 2 using user terminal 4a includes a user identification label associated with attribute label LA corresponding to company A. Similarly, the user information referenced when a user using user terminal 4b logs in to the information processing device 2 using user terminal 4b includes a user identification label associated with attribute label LB corresponding to company B.

[0021] In this embodiment, the user identification label matches the attribute label. Therefore, in FIG. 1, the user identification label assigned to the user terminal 4a used by a user belonging to company A is denoted as LA, and the user identification label assigned to the user terminal 4b used by a user belonging to company B is denoted as LB. Note that, although the example in FIG. 1 illustrates two exclusive databases 32a, 32b and two user terminals 4a, 4b, there may be two or more exclusive databases and two or more user terminals. Furthermore, multiple user terminals operated by multiple users belonging to one company may be able to access the information processing device 2. The information processing device 2 may be able to access multiple exclusive databases belonging to one company.

[0022] In this way, the multiple proprietary databases 32a, 32b are databases managed by organizations to which users who use user terminals 4a, 4b belong, and whose user information includes user identification labels LA, LB associated with attribute labels LA, LB assigned to the proprietary databases 32a, 32b. Organizations include organizations such as companies, business establishments, membership groups, etc. The information processing device 2 is managed by a product provider or a third party. The information processing device 2 and the proprietary databases 32a, 32b are communicatively connected via a predetermined network N such as the Internet. The information processing device 2 and the common database 31 may be connected by wire or via the predetermined network N.

[0023] For example, the attribute label includes the domain of the corresponding proprietary database 32a, 32b. In this case, the attribute label may include the domain of the organization managed by the proprietary database 32a, 32b. In this embodiment, the user identification label may include the domain of the organization to which the user belongs. If the attribute label is the domain of the organization managed by the proprietary database 32a, 32b and the user identification label is the domain of the organization to which the user belongs, the user identification label matches the attribute label.

[0024] Fig. 2 is a block diagram showing the functional configuration of the information processing device shown in Fig. 1. As shown in Fig. 2, the processing circuit 20 of the information processing device 2 has, as functional blocks, a user authentication unit 22, a question analysis unit 23, a query generation unit 24, a search execution unit 25, and an answer generation unit 26. Each of these functional blocks is also considered to be a processing circuit. The example of Fig. 2 illustrates a case where a user terminal 4a, whose user information includes a user identification label LA, accesses the information processing device 2.

[0025] 3 is a flowchart showing an example of the flow of information provision processing in the information processing device of this embodiment. The user authentication unit 22 authenticates access from the user terminal 4a and permits login (step S1). The user authentication unit 22 receives user information transmitted from the user terminal 4a and performs authentication based on the user information. The user authentication unit 22 acquires a user identification label LA included in the user information.

[0026] The question analysis unit 23 receives the question content input based on the operation of the user terminal 4a and analyzes the question content (step S2). The question content includes the product number, the defective part, the symptoms, etc. For example, the question analysis unit 23 prompts the user to input the question content in a Q&A format using a chatbot or the like. Alternatively, the question analysis unit 23 may receive a text sentence regarding the defect sent from the user terminal 4a.

[0027] The query generation unit 24 generates a search query for a predetermined database in the database group 3 using a large-scale language model from the content of the question (step S3). The query generation unit 24 acquires the user identification label LA of the user who is the questioner, and determines a predetermined database to which the search query is to be sent based on the user identification label LA (step S4). Of the multiple occupied databases 32a and 32b, the query generation unit 24 determines the occupied database 32a to which the attribute label LA associated with the user identification label LA is assigned, and the common database 31, as the predetermined databases.

[0028] The search execution unit 25 acquires the generated search query and information on the predetermined database to be searched, and transmits the search query to the exclusive database 32a determined as the predetermined database and the common database 31 (step S5). The search query is not transmitted to the exclusive database 32b to which the attribute label LB is assigned that is not associated with the user identification label LA of the user who is the questioner.

[0029] The answer generation unit 26 obtains the search results from the predetermined database 31, 32a to which the search query was sent, and generates an answer to the question from the search results using a large-scale language model (step S6). In this case, the search execution unit 25 or the answer generation unit 26 may analyze the search results and extract search results necessary for answer generation. The answer generation unit 26 outputs the generated answer to the user terminal 4a (step S7).

[0030] After generating the answer, the processing circuitry 20 can also train a large-scale language model using the answer and the question content.

[0031] Fig. 4 is a block diagram showing the functional configuration of the information processing device shown in Fig. 1 when accessed from a user terminal different from that shown in Fig. 2. The example of Fig. 4 illustrates a case where a user terminal 4b, whose user information includes a user identification label LB, accesses the information processing device 2. In Fig. 4, functional blocks 22, 23, 24, 25, and 26 in the processing circuit 20 are the same as those in Fig. 2. The example of Fig. 4 differs from Fig. 2 in that the user terminal accessing the information processing device 2 is different and that the predetermined database accessed by the information processing device 2 is different.

[0032] That is, the user information transmitted from the user terminal 4b to the information processing device 2 includes the user identification label LB, and the user authentication unit 22 acquires the user identification label LB included in the user information. Therefore, the query generation unit 24 determines, as predetermined databases, the exclusive database 32b, to which the attribute label LB associated with the user identification label LB is assigned, and the common database 31, from among the multiple exclusive databases 32a and 32b.

[0033] The search execution unit 25 acquires the generated search query and information on the predetermined database to be searched, and transmits the search query to the exclusive database 32b determined as the predetermined database and the common database 31. The search query is not transmitted to the exclusive database 32a to which the attribute label LA is assigned that is not associated with the user identification label LB of the user who is the questioner.

[0034] In this way, at least a part of the predetermined database to be searched differs depending on the user identification labels LA, LB included in the user information of the user terminals 4a, 4b, so even if the exact same question is entered in the user terminals 4a and 4b, the search results may differ. Therefore, the answers generated by the answer generator 26 may also differ depending on the user identification labels LA, LB included in the user information of the user terminals 4a, 4b.

[0035] According to this embodiment, the databases to be searched include multiple proprietary databases 32a, 32b with different attribute labels LA, LB, and the user identification labels included in the user information of the user accessing the information processing device 2 are associated with the attribute labels LA, LB, so that the database to be searched can be determined depending on the user. This allows the information processing device 2 to seamlessly access the unique databases that the user has permission to access in response to the user's question, thereby increasing the accuracy of answers to the user's question. This makes it possible to provide countermeasures for product defects that are tailored to the user's usage status, etc., without relying on a specific person in charge.

[0036] Furthermore, since access from the information processing device 2 is not permitted to the proprietary database that is not associated with the user identification labels LA and LB transmitted from the user terminals 4a and 4b, it is possible to prevent confidential information from leaking from the proprietary database to users who are not permitted to access it.

[0037] Furthermore, in this embodiment, the query generation unit 24 determines the common database 31 as the predetermined database to which a search query is sent, regardless of the user identification labels LA and LB included in the user information. This eliminates the need to store data that is not specific to a user in the exclusive databases 32a and 32b, which are accessed differently for each user, thereby suppressing an increase in the overall capacity of the database group 3. Furthermore, by separately constructing the common database 31 and the exclusive databases 32a and 32b, the database group 3 can be constructed efficiently and easily.

[0038] In this case, the provider of the search system 1 manages user-independent data as the common database 31, and the proprietary databases 32a and 32b are managed by someone other than the provider of the search system 1, such as a product user, thereby reducing the capacity of the databases managed by the provider of the search system 1. Also, databases that can be searched by the information processing device 2 can be easily integrated and separated. Furthermore, confidentiality is improved because confidential information contained in the proprietary databases 32a and 32b does not need to be provided to the provider of the search system 1. Furthermore, because searches can be performed using databases that contain confidential information accessible to the user, more accurate answers tailored to the user can be provided to the user's questions.

[0039] In this embodiment, the attribute label includes the domain of the corresponding occupancy database 32a, 32b. Therefore, if the domain of the occupancy database 32a, 32b is registered in the information processing device 2, there is no need to register the attribute label separately, which can save time and effort.

[0040] Although the embodiments of the present disclosure have been described above, the present disclosure is not limited to the above-described embodiments, and various improvements, changes, and modifications are possible.

[0041] [Other embodiments] For example, in the above embodiment, the proprietary databases 32a, 32b are entirely managed by an organization such as a company associated with the assigned attribute labels LA, LB, but this is not limiting. For example, the provider of the product or the provider of the search system 1 may manage all of the proprietary databases 32a, 32b, or some of the proprietary databases 32a, 32b may be managed by an organization associated with the attribute labels LA, LB, and the rest may be managed by the provider of the search system 1.

[0042] For example, data contained in the proprietary databases 32a and 32b managed by the product provider may include past maintenance information, system configuration information, and the like, recorded for each user of the product. Furthermore, the proprietary databases 32a and 32b may be databases managed by a system integrator that performs system integration to apply a robot, which is a product, to a system. Furthermore, the provider of the search system 1 and the administrator of the common database 31 may be different persons. In this case, the provider of the search system 1 does not need to have a database that it manages.

[0043] Furthermore, in the above embodiment, an example was shown in which the user identification label and the attribute label are the same, but the user identification label and the attribute label may be different. For example, the user identification label may be the name of the organization to which the user belongs, and the attribute label may be the domain of the corresponding proprietary databases 32a and 32b. When the user identification label and the attribute label are different, the memory 21 stores the correspondence between the user identification label and the attribute label. This allows the processing circuit 20 to identify the proprietary databases 32a and 32b to be searched by the information processing device 2 based on the content of the user's question from the user identification label included in the user information.

[0044] In the above embodiment, the user identification label and the attribute label are associated with each other so that the proprietary databases 32a and 32b searched by the information processing device 2 in response to a user's question are the proprietary databases 32a and 32b managed by an organization such as a company to which the user who is the questioner belongs. However, the present invention is not limited to this. For example, the proprietary databases searched by the information processing device 2 may differ between multiple users who belong to the same company.

[0045] For example, the proprietary databases that the information processing device 2 searches may differ depending on the position of the user accessing the information processing device 2. For example, if two proprietary databases are managed by a certain company and a first user who is a manager of the company accesses the information processing device 2, the information processing device 2 may include the two proprietary databases in the search targets, whereas if a second user who is not a manager of the company accesses the information processing device 2, the information processing device 2 may include one of the two proprietary databases in the search targets but not the other proprietary database. Alternatively, the proprietary databases that the information processing device 2 searches may differ depending on the department of the user who accesses the information processing device 2.

[0046] Furthermore, in the above embodiment, the search system 1 that provides information on how to deal with product defects has been exemplified, but the search system 1 to which the information processing device 2 of the present disclosure is applied is not limited to this. For example, the information processing device 2 of the present disclosure can be applied to knowledge search systems in a wide range of fields, such as search systems for dealing with various business problems.

[0047] Furthermore, in the above embodiment, an example was given in which the query generation unit 24 determines the database to be searched, but as long as the processing circuit 20 of the information processing device 2 determines the database to be searched, other functional blocks such as the user authentication unit 22 or the search execution unit 25 may perform the function.

[0048] Furthermore, in the above embodiment, an example has been shown in which the information processing device 2 is configured as a single device including the processing circuitry 20 and the storage unit 21 in which the large-scale language model is stored, but the processing circuitry 20 and the storage unit 21 may be configured as separate devices. Some of the functional blocks of the processing circuitry 20 may also be configured as separate devices. In other words, the information processing device 2 may be configured from multiple processing circuits.

[0049] For example, a processing circuit for user authentication and determining a specific database to be searched and a processing circuit for generating a search query from the question content using a large-scale language model and generating an answer from the search results may be configured as separate processing circuits.

[0050] Summary of this disclosure [Item 1] An information processing device according to one aspect of the present disclosure is an information processing device that can access multiple databases and communicate with multiple user terminals, and is equipped with a memory device in which a large-scale language model is stored, and a processing circuit, wherein the processing circuit performs authentication based on user information sent from the user terminal, accepts a question input based on operation of the user terminal, generates a search query for a specified database among the multiple databases using the large-scale language model from the question content, sends the search query to the specified database, generates an answer to the question from search results in the specified database using the large-scale language model, and outputs the answer to the user terminal, wherein the multiple databases include a proprietary database to which an attribute label is assigned, and the user information includes a user identification label associated with the user using the user terminal, and the processing circuit determines the proprietary database among the multiple databases to which the attribute label corresponding to the user identification label included in the user information is assigned as the specified database to which the search query is sent.

[0051] According to the above configuration, the database to be searched includes a plurality of proprietary databases with different attribute labels, and the user identification labels included in the user information of the user accessing the information processing device are associated with the attribute labels, so that the database to be searched can be determined according to the user. This allows the information processing device to seamlessly access the unique databases that the user is authorized to access in response to the content of the user's question, thereby increasing the accuracy of answers to the user's questions.

[0052] [Item 2] In the information processing device of item 1, the attribute label may include the domain of each database. This eliminates the need to register a separate attribute label by simply registering the domain of the proprietary database in the information processing device 2, thereby saving time and effort.

[0053] [Item 3] In the information processing device of item 1 or 2, the plurality of databases may include a common database to which no attribute label is assigned, and the processing circuit may determine the common database as the predetermined database to which the search query is sent, regardless of the user identification label included in the user information. By constructing the common database and the exclusive database separately, the capacity of the exclusive database can be reduced, and a group of databases including these can be constructed efficiently and easily.

[0054] [Item 4] In the information processing device of item 3, the question may include a question about how to deal with a defect that has occurred in a specific product, the common database may include a database of the product's manual or a database of defect information for the product held by the product provider, and the proprietary database may include a database of defect information for the product held by a user of the product or a database of product usage status by the user of the product. By storing data on product defect information and usage status that the user does not want others to know in the proprietary database, it becomes possible to search using a database containing confidential information accessible to the user. Therefore, it is possible to provide more accurate answers to the user's questions that are tailored to the user.

[0055] [Item 5] In the information processing device of any one of items 1 to 4, the proprietary database may be a database managed by an organization to which a user who uses a user terminal whose user information includes the user identification label associated with the attribute label assigned to the proprietary database belongs. This eliminates the need to provide confidential information included in the proprietary database to a provider of a search system, thereby improving confidentiality.

[0056] [Item 6] Another aspect of the present disclosure is an information provision method that generates an answer to a question sent from a user terminal and provides the answer, performing authentication based on user information sent from the user terminal, accepting the question input based on operation of the user terminal, generating a search query to a predetermined database from a plurality of databases using a large-scale language model from the question, sending the search query to the predetermined database, generating an answer to the question from search results in the predetermined database using the large-scale language model, and outputting the answer to the user terminal, wherein the plurality of databases include a proprietary database to which an attribute label is assigned, the user information includes a user identification label associated with a user using the user terminal, and determining the proprietary database from the plurality of databases to which the attribute label corresponding to the user identification label included in the user information is assigned as the predetermined database to which the search query is sent. [Explanation of symbols]

[0057] 2. Information processing equipment 4a, 4b User terminal 20 Processing circuit 21 Memory 31 Common Database 32a, 32b Proprietary Database

Claims

1. An information processing device that can access a plurality of databases and communicate with a plurality of user terminals, a memory in which a large-scale language model is stored; processing circuitry; The processing circuitry performing authentication based on user information transmitted from the user terminal; Accepting a question input based on an operation of the user terminal; generating a search query for a predetermined database among the plurality of databases from the content of the question using the large-scale language model; sending the search query to the predetermined database; generating an answer to the question using the large-scale language model from a search result in the predetermined database; An answer is output to the user terminal, the plurality of databases includes a proprietary database with attribute labels; the user information includes a user identification label associated with a user who uses the user terminal; The processing circuit determines the dedicated database among the plurality of databases to which the attribute label corresponding to the user identification label included in the user information is assigned as the specified database to which the search query is sent.

2. The information processing apparatus according to claim 1 , wherein the attribute labels include a domain of each database.

3. the plurality of databases includes a common database to which no attribute labels are assigned; The information processing device according to claim 1 , wherein the processing circuit determines the common database as the predetermined database to which the search query is to be sent, regardless of the user identification label included in the user information.

4. The content of the question includes a question regarding how to deal with a defect that has occurred in a predetermined product, the common database includes a database relating to the product manual or a database relating to defect information of the product held by the product provider, The information processing apparatus according to claim 3 , wherein the proprietary database includes a database relating to defect information of the product held by a user of the product, or a database relating to a usage status of the product by the user of the product.

5. 3. The information processing device according to claim 1, wherein the proprietary database is a database managed by an organization to which a user who uses a user terminal whose user information includes the user identification label associated with the attribute label assigned to the proprietary database belongs.

6. An information providing method for generating an answer to a question sent from a user terminal and providing the answer, comprising: Authentication is performed based on user information sent from the user terminal, Accepting a question input based on an operation of the user terminal; generating a search query for a predetermined database among a plurality of databases from the content of the question using a large-scale language model; sending the search query to the predetermined database; generating an answer to the question using the large-scale language model from a search result in the predetermined database; An answer is output to the user terminal, the plurality of databases includes a proprietary database with attribute labels; the user information includes a user identification label associated with a user who uses the user terminal; An information providing method, wherein the dedicated database among the plurality of databases to which the attribute label corresponding to the user identification label included in the user information is assigned is determined as the specified database to which the search query is sent.

Images