Information provision system, information provision method, and information provision program
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- POCKET SIGN CO LTD
- Filing Date
- 2026-06-18
- Publication Date
- 2026-08-03
Smart Images
Figure 0007898798000001_ABST
Abstract
Description
Technical Field
[0001] Embodiments of the present invention relate to an information providing system, an information providing method, and an information providing program.
Background Art
[0002] Conventionally, there is known a system that allows a user to simply search for an appropriate support system of an administrative agency and understand the procedure method by merely inputting the information they seek. Also, a technique using a generative AI for such searches is known. On the other hand, administrative agencies are using My Number cards to provide various services to residents.
Prior Art Documents
Patent Documents
[0003]
Patent Document 1
Patent Document 2
Non-Patent Documents
[0004]
Non-Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0005] There are many different types of free information services that use AI generation. Government agencies also provide information services funded by taxpayers on their websites. However, if one user uses a large amount of AI generation, the AI's resources become monopolized by that user, creating an unfair situation for other users who do not use it much. In addition, there are costs associated with operating AI generation, so there is a need to restrict excessive use by users in some way while ensuring fairness.
[0006] One example of a problem that this invention aims to solve is ensuring fairness in the use of a learning model by controlling the amount of usage according to the user. [Means for solving the problem]
[0007] The invention according to this embodiment is, An information provision system in which a specific organization provides information to external users, A user authentication unit that authenticates the user's identity using a medium capable of verifying identity, An identification information assignment unit that assigns identification information to the user based on the aforementioned authentication, A usage control unit that controls the usage of a machine learning model that provides the user with output information output in response to the user's request input based on the aforementioned identification information, It is an information provision system equipped with [features / equipment].
[0008] The invention according to this embodiment is, One or more computers that implement an information provision system in which a specific organization provides information to external users, An identity verification process that authenticates the user's identity using a media capable of verifying identity, Based on the aforementioned authentication of the user, an identification information assignment process is performed to assign identification information to the user, A usage control process that controls the usage of a machine learning model that provides the user with output information output in response to the user's request input, based on the aforementioned identification information. An information providing method that executes
[0009] The invention according to this embodiment On one or more computers that realize an information providing system in which a specific organization provides information to an external user A personal authentication process for authenticating the user using a personal confirmation possible medium An identification information assigning process for assigning identification information to the user based on the personal authentication A usage amount control process for controlling the usage amount of a learned learning model that provides the user with output information output in response to an input of the user's request based on the identification information An information providing program that causes the above to be executed
Effect of the Invention
[0010] According to this embodiment, by controlling the usage amount according to the user, fairness regarding the use of the learning model can be ensured.
Brief Description of the Drawings
[0011] [Figure 1] It is a system configuration diagram showing an overall view of the information providing system. [Figure 2] It is a block diagram showing the information providing system. [Figure 3] It is a block diagram showing the hardware configuration of a computer. [Figure 4] It is an explanatory diagram showing the content registered in the database. [Figure 5] It is a flowchart showing the information providing process. [Figure 6] It is a flowchart showing the learning model usage process.
Mode for Carrying Out the Invention
[0012] Hereinafter, the present invention will be described through embodiments of the invention, but the invention according to the claims is not limited to the following embodiments. Also, not all of the configurations described in the embodiments are essential as means for solving the problems. For the sake of clarity of explanation, the following description and drawings have been appropriately omitted and simplified. In each drawing, the same reference numerals are assigned to the same elements, and duplicate explanations are omitted as necessary.
[0013] (Information providing system 1) Referring to FIG. 1, the general configuration of the information providing system 1 will be described. FIG. 1 is a system configuration diagram showing an overall view of the information providing system 1. An information providing method is implemented using this information providing system 1. Further, the information providing method of the present embodiment is realized by causing a computer to execute an information providing program.
[0014] The information providing system 1 of the present embodiment is a system in which a specific organization provides predetermined information to external users using artificial intelligence such as a generative AI. "External users" include, for example, users who do not belong to a specific organization. If the specific organization is a city hall, the "external users" may be general users other than the city hall staff. The specific organization is exemplified by an administrative agency G. Note that the specific organization may be other organizations, for example, private organizations (private companies) or public organizations. The information provided by the information providing system 1 may be information that is widely and generally publicly available on the Internet or the like, or may be non-public information possessed only by a specific organization. In the following description, the information provided by the information providing system 1 is referred to as output information. The output information includes answers generated by the generative AI, items searched by the generative AI, and other general information.
[0015] This information provision system 1 uses artificial intelligence to provide output information, and controls the amount of artificial intelligence used according to the user U. Generally, systems built with neural networks such as large-scale language models incur significant costs for implementation and operation. For example, government agencies G may use tax money to operate large-scale language models and provide information using them free of charge. In such cases, there is a need to ensure fairness while limiting excessive use by a few users U. Therefore, information provision system 1 uses My Number cards or similar to authenticate the identity of user U (identity verification authentication), counts the amount of large-scale language model used by user U, and restricts its use when a predetermined amount is reached.
[0016] The information provision system 1 primarily comprises at least an information provision server 10. The information provision server 10 is a computer managed and operated by administrative agency G, which is a specific organization. Furthermore, the information provision system 1 includes a certificate verification server 20, an administrator terminal 30, a user terminal 40, a medium capable of verifying identity 41, and a resident information management platform 50.
[0017] The information provision server 10 is located on the cloud. For example, the information provision server 10 is connected to a predetermined network N, such as the internet, in a communicative manner. Note that network N is not limited to the internet; it may also be a LAN (Local Area Network), WAN (Wide Area Network), or mobile communication network.
[0018] The certificate verification server 20, resident information management platform 50, administrator terminal 30, and user terminal 40, all acting as computers, are connected to the information provision server 10 via network N.
[0019] The certificate verification server 20 is a computer located in the cloud. The certificate verification server 20 verifies the validity of the digital certificate stored in the identity verification medium 41. The certificate verification server 20 is provided by the information provision server 10 to perform identity verification authentication of user U using the public personal authentication service. For example, if the digital certificate stored in the identity verification medium 41 is valid (if identity verification authentication is successful), the verified user information of user U can be registered with the information provision server 10.
[0020] Furthermore, the information provision server 10 may have the resident information management platform 50 perform identity verification authentication (resident verification authentication) of user U using the public personal authentication service. For example, the information provision server 10 may be allowed to access the resident information management platform 50 if the electronic certificate stored in the identity verification medium 41 is valid.
[0021] The information provision server 10 intervenes between the identity verification medium 41 owned by user U and the public personal authentication service, and enables the public personal authentication service to verify the identity of user U. Alternatively, the information provision server 10 may verify the identity of user U using a card-alternative electronic record.
[0022] Card-alternative electronic records are stored on designated devices such as smartphones. Card-alternative electronic records are essentially mobile documents (mdocs) that store My Number Card information on smartphones, allowing for identity verification without the need for a physical card.
[0023] The certificate verification server 20 is responsible for at least part of the function of verifying the electronic certificates obtained from the identity verification medium 41 in the public personal authentication service. Specifically, the certificate verification server 20 manages the revocation information of the electronic certificates.
[0024] The certificate verification server 20 may, for example, perform authentication for the public personal authentication service operated by the Japan Local Government Information Systems Organization (J-LIS). The public personal authentication service is a means of identity verification used when performing administrative procedures such as online applications and notifications via the internet. In addition, the certificate verification server 20 may also include, for example, a signature verifier device that verifies the electronic signature of the public personal authentication service based on the revocation information received from J-LIS.
[0025] The resident information management platform 50 is a computer. When the information provision server 10 accesses the resident information management platform 50, it performs identity verification authentication (resident verification authentication) of user U.
[0026] The Resident Information Management Platform 50 is a platform for national or local governments to manage resident information. The Resident Information Management Platform 50 networks the four basic pieces of information from the Basic Resident Register (name, address, date of birth, and gender), which form the basis of various administrative services, as well as individual numbers, resident registration codes, and information on changes to these. The Resident Information Management Platform 50 may, for example, include the functions of the Digital Agency's My Number Portal API, or it may include the internal systems of local governments.
[0027] The Resident Information Management Platform 50 manages various types of information, including residents' taxes, income, households, vaccinations, and pensions. The Resident Information Management Platform 50 accepts access from residents who have successfully undergone identity verification using the Public Personal Authentication Service. Typically, the Resident Information Management Platform 50 plays a role in providing administrative information to residents through the My Number Portal, an administrative service operated by the government. In other words, the Resident Information Management Platform 50 may be envisioned as a resident information management infrastructure of the national or local government (local public body), including the My Number Portal API.
[0028] For example, consider a scenario where user U, who has successfully completed identity verification, accesses the My Number Portal and subsequently wishes to obtain family register information via the My Number Portal. User U performs resident verification authentication to obtain family register information. If this resident verification authentication is successful, the Family Register Information Linkage System, which is under the jurisdiction of the Ministry of Justice, receives the request from the My Number Portal and, through this Family Register Information Linkage System, can obtain the family register information from the municipal database of the city hall where user U's permanent domicile is located.
[0029] Furthermore, the resident information management platform 50 may be an internal system operated by a local government. The resident information management platform 50 may also be responsible for accepting access using the serial number of user U obtained through the public personal authentication service, four pieces of basic information, and derived information thereof, and for providing information about this user U held by the local government.
[0030] The administrator terminal 30 is, for example, a computer such as a personal computer owned by a designated administrator M. Administrator M is, for example, an employee within the administrative agency G who has been granted the prescribed authority regarding information provision. Administrator M uses the administrator terminal 30 to access the information provision server 10 and makes decisions to approve the provision of prescribed information provided from the information provision server 10 to user U, or to edit the information.
[0031] The user terminal 40 is, for example, a computer such as a personal computer owned by user U. The user terminal 40 may also be a smartphone or a tablet computer. User U has the user terminal 40 read their own personal identification medium 41. The information provision server 10 then uses the certificate verification server 20 to authenticate user U's identity.
[0032] The user terminal 40 includes a media reader (not shown) that reads the identity verification medium 41. This media reader reads the electronic certificate embedded in the integrated circuit built into the identity verification medium 41. The media reader also has, for example, a short-range wireless communication function that conforms to a predetermined standard. The media reader may be located outside the user terminal 40, or it may be connected to the user terminal 40 when in use.
[0033] Furthermore, the user terminal 40 may have a function for card-alternative electronic recording that replaces the identity verification medium 41. If the user terminal 40 has a card-alternative electronic recording function, the configuration of the medium reader unit may be omitted. Also, user U does not need to possess the identity verification medium 41.
[0034] The identity verification medium 41 is, for example, a My Number Card issued by a public institution. The identity verification medium 41 can be any medium that allows user U's identity verification by the public personal authentication service, such as a My Number Card integrated with a driver's license or a specific residence card. The identity verification medium 41 does not have to be an integrated My Number Card, such as a driver's license, passport, or residence card. These mediums are equipped with an IC chip that stores electronic certificates, etc.
[0035] Furthermore, the My Number Card is an IC card that can be used as an identification document for identity verification, as well as for various services such as local government services and electronic applications using electronic certificates, such as e-Tax.
[0036] The My Number Card has the My Number (individual number), which is legally limited in the scope of its use, printed on its surface. Furthermore, the My Number Card is equipped with an IC chip that stores electronic certificates and applications (APs) for the public personal authentication service, which can be widely used by private businesses as well. This IC chip stores digital signature certificates and user authentication certificates, among others. Digital signature certificates are used when creating and sending electronically signed electronic documents over the internet. User authentication certificates are used when logging into various service sites provided by the government and private sector. Users of the My Number Card can use these electronic certificates after being authenticated by the authentication authority by entering the PIN set for each electronic certificate. The identity verification medium 41 may include at least one of the following: the My Number Card, a smartphone equipped with a prescribed electronic certificate (so-called smartphone JPKI), and a smartphone on which a card substitute electromagnetic record is recorded.
[0037] (Block diagram) Next, a block diagram showing the information provision system 1 will be explained with reference to Figure 2. Note that each component of the information provision system 1 represents a functional block, not a hardware-level configuration. Each component is implemented by any combination of hardware and software, centering on the CPU, memory, programs loaded into memory, storage media such as a hard disk for storing those programs, and a network connection interface. Furthermore, there are various modifications to the implementation method and apparatus.
[0038] In this embodiment, we illustrate a configuration in which each component of the information provision system 1 is provided on the information provision server 10.
[0039] The information provision server 10 includes a user authentication unit 101, an identification information assignment unit 102, a model switching unit 103, an audio / video processing unit 104, a usage control unit 105, a restriction release unit 106, a provision management unit 107, and an information provision unit 108. These are realized by programs stored in memory or an HDD (Hard Disk Drive) being executed by a CPU (Central Processing Unit).
[0040] Furthermore, the information provision unit 108 includes multiple types of learning models 109 and an extension generation unit 110. The learning models 109 are managed by the administrative agency G (a specific agency). Note that these learning models 109 do not necessarily have to be learning models 109 owned by the administrative agency G; for example, the administrative agency G may use learning models 109 owned by another company through a service provided by that company, paying a fee for such a service via API integration. This embodiment includes such API integration. In other words, the statement "the learning models 109 are managed by the administrative agency G (a specific agency)" includes the API integration form. Also, the learning models 109 may be tuned specifically for the administrative agency G in order to be managed by the administrative agency G.
[0041] Furthermore, the information provision server 10 includes a predetermined database 111. This database 111 is a collection of information that is stored in memory, HDD, or cloud computing resources and organized so that it can be searched or stored. The database 111 stores the data necessary for the learning model 109 to generate answers.
[0042] The identity verification unit 101 performs identity verification (identity confirmation authentication) of user U using the identity verification medium 41. For example, the identity verification unit 101 performs identity verification of user U using the public personal authentication service on the identity verification medium 41. Note that the functions of the identity verification unit 101 may also be located in the certificate verification server 20.
[0043] The identification information assignment unit 102 assigns a unique user ID to user U as unique identification information based on the user authentication. For example, when the aforementioned user authentication is successful, the identification information assignment unit 102 may assign a unique user ID using information indicating that the user authentication was successful. The unique user ID is the identification information of user U that is generated through authentication by the public personal authentication service using the identity verification medium 41 issued by a public institution. For example, even if the name, address, etc., listed on the My Number Card or electronic certificate are changed, the unique user ID is identification information that does not change.
[0044] Furthermore, when the identification information assignment unit 102 assigns identification information to the same user U whose information on the identity verification medium 41 has been updated, based on identity authentication (identity verification authentication) using the identity verification medium 41 again, it assigns the same identification information as the previously assigned identification information.
[0045] For example, consider a scenario where the electronic certificate used for identity verification has changed (been updated) due to user U moving, etc. In this case, the identification information assignment unit 102 refers to the history of changes to the electronic certificate's serial number, or the connections between the serial numbers of multiple types of electronic certificates, using the serial number linking function provided by J-LIS. If the identification information assignment unit 102 determines that the user is the same person who was previously registered with the information provision server 10, it returns the same user-specific ID. If the identification information assignment unit 102 determines that the user is not the same person who was previously registered, it issues a new user-specific ID. In this way, the identification information assignment unit 102 has the function of determining whether user U has been previously authenticated and registered by the public personal authentication service using the identity verification medium 41. In this manner, the identification information assignment unit 102 can assign one piece of identification information to a single user U that remains unchanged throughout their life. In this embodiment, the term "assign" includes the meanings of "issue" or "return".
[0046] The functionality of the identification information assignment unit 102 may also be located in the certificate verification server 20. The information provision server 10 may access the certificate verification server 20 when assigning a user-specific ID. Alternatively, the information provision server 10 may use the user-specific ID issued by the certificate verification server 20 itself, or it may use its own user ID associated with that user-specific ID.
[0047] The model switching unit 103 switches the type of learning model 109 based on attribute information indicating the user U's attributes, which can be obtained from a system managed by a public institution such as the resident information management platform 50, and outputs output information from the learning model 109. In this way, the optimal learning model 109 can be used according to the user U.
[0048] The learning model 109 in this embodiment outputs text information or audio or video output information in response to a request (question) input from user U in the form of text information. Furthermore, the learning model 109 outputs audio or video output information in the form of at least one of the audio or video (including video and still images) input from user U in the form of at least one of the audio or video. Here, audio or video input / output places a significantly greater processing load on the learning model 109 compared to text information input / output. According to this embodiment, the processing load on the learning model 109 can be reduced by limiting the amount of data used by user U.
[0049] The audio-video processing unit 104 performs processing to enable communication between the user U and the learning model 109 using at least one of either audio or video. In this way, communication can be performed using at least one of either audio or video, in addition to communication using text information.
[0050] The usage control unit 105 controls the usage of the machine-learned model 109, which provides user U with output information output in response to user U's request input, based on identification information (user-specific ID). In this way, fairness regarding the use of the machine-learned model 109 can be ensured by controlling the usage according to user U.
[0051] Furthermore, if the specified institution is a public institution, the usage control unit 105 increases or decreases the usage of the learning model 109 based on attribute information indicating the attributes of user U that can be obtained by the system managed by the public institution. In this embodiment, usage amounts include, for example, the number of uses, usage time, number of sessions used, amount of data used, number of tokens used, and number of words used.
[0052] The usage control unit 105 may increase or decrease the usage amount based on attribute information obtainable through user authentication (identity verification medium 41), or based on user U attribute information obtainable through a system managed by a public institution. Furthermore, the usage control unit 105 may increase or decrease the usage amount by taking both into consideration. In addition, the increase or decrease in usage amount includes the case where the usage amount is zero. In other words, the increase or decrease in usage amount includes the mode of permission or denial of use.
[0053] For example, if the address obtained through user authentication is in Miyagi Prefecture, the usage control unit 105 may set the number of uses (usage) to 10 if user U is a resident of Miyagi Prefecture, and set the number of uses (usage) to 5 if user U is not a resident of Miyagi Prefecture.
[0054] For example, if the public institution is a city hall, the usage control unit 105 may access a non-public city hall database managed by the city hall and obtain user U's attribute information from it. In this way, fairness can be ensured while providing each user U with an appropriate usage amount using the learning model 109.
[0055] For example, the usage control unit 105 controls the usage of the learning model 109 according to the address, age, hometown tax donation status, high taxpayer status, high-income household status, tax-exempt household status, child-rearing household status, place of residence, place of tax payment, place of origin, nationality, etc. For example, a user U who is a hometown tax donor, a high taxpayer, a high-income household, a tax-exempt household, or a child-rearing household may have increased usage compared to a typical user U.
[0056] For example, if the specified institution is the Miyagi Prefectural Government, the usage control unit 105 may set the number of uses (usage) to 10 if user U is a resident of Miyagi Prefecture, and set the number of uses (usage) to 5 if user U is not a resident of Miyagi Prefecture.
[0057] In this embodiment, public institutions include, for example, organizations operated and managed by the national or local government, such as the Japan Local Government Information System Organization (J-LIS), the Digital Agency, the Immigration Services Agency, and local governments.
[0058] Information regarding address and age includes, for example, information obtainable through the Public Personal Authentication Service or information obtainable through the Card Surface Information Input Assistance Application (Card Surface AP). Here, Card Surface AP is an application for reading information recorded in the IC chip of the My Number Card, which is written on the surface of the card (front or back), as text data.
[0059] Tax information includes, for example, information obtainable through the My Number Portal API (Resident Information Management Platform 50). Information on hometown tax donations, high taxpayers, high-income households, tax-exempt households, and child-rearing households includes, for example, information obtainable through the city hall system. Information on foreign residents includes, for example, information obtainable through residence cards.
[0060] The restriction release unit 106 releases the usage limit based on at least one of predetermined conditions or an operation by the administrator M of the administrative agency G (specific agency). For example, if the usage of a designated user U is restricted, the administrator M can manually release the restriction. The administrator M can also set conditions in advance for releasing the usage limit. For example, the administrator M can set it so that the restriction is released when an earthquake occurs. In this way, the usage limit can be automatically released in emergencies such as when an earthquake occurs.
[0061] The provision management unit 107 makes approval decisions for provision or edits based on the output information, based on the operations of administrator M. Before the output information is provided to user U, administrator M can review it, and administrator M can approve the provision of the output information or edit it based on the output information. The provision management unit 107 receives the operation of approving the provision of output information or editing it from administrator M. In this way, administrator M inside administrative agency G can scrutinize the information provided, and more accurate information can be provided to user U.
[0062] The information provision unit 108 provides output information generated using the machine learning model 109 in response to user U's request. In this way, even if user U's request is ambiguous, the machine learning model 109 can be used to answer user U's request. User U's request is a question to the machine learning model 109. The request may be a predetermined prompt to be input to the machine learning model 109, or it may be a predetermined file.
[0063] The information provision unit 108 inputs input data, including the user U's request (question), into the learning model 109, and causes it to output output data, including the answer corresponding to the request. The information provision unit 108 then transmits the answer generated by the learning model 109 to the user terminal 40.
[0064] The learning model 109 is, for example, a large-scale language model. A large-scale language model is pre-trained on general characteristics of question and answer trends. A large-scale language model is a machine learning model that learns from large amounts of text data using deep learning, enabling it to understand and generate natural-sounding text like a human. Large-scale language models are also called LLMs (large language models). Note that instead of a large-scale language model, a small-scale language model (SLM) or similar may be used as the learning model 109.
[0065] The learning model 109 is developed by various companies, and different types are developed depending on the application. The information provision unit 108 is equipped with multiple types of learning models 109. In this way, it can generate an answer using the optimal learning model 109 according to the user U's request. However, the information provision unit 108 may be equipped with only one type of learning model 109.
[0066] The augmented generation unit 110 has the function of storing information related to various types of information provided to user U and incorporating it into the learning model 109. This improves the accuracy of the response results. The augmented generation unit 110 also has a RAG (Retrieval Augmented Generation) function. This allows the output information output from the learning model 109 to be augmented according to various requests from user U. The information stored by the augmented generation unit 110 may be customized according to user U's attribute information.
[0067] For example, the extended generation unit 110 searches the database 111 for information related to the user U's request (question) based on the user U's attribute information. The extended generation unit 110 adds the relevant information found in the search to the learning model 109. In other words, the extended generation unit 110 extends the functionality of the learning model 109. In this way, the learning model 109 can generate an accurate answer that is close to the answer the user U is looking for by combining the extended information with the learned knowledge. Note that one extended generation unit 110 may be set up for each user U.
[0068] Furthermore, the learning model 109 may be pre-trained to not answer inappropriate questions. The learning model 109 may also be controlled by the extended generation unit 110 or input prompts to prevent it from answering inappropriate questions. Additionally, the learning model 109 may be trained to acquire information specific to administrative agency G, such as the latest minutes of council meetings at administrative agency G. This allows for the rapid provision of information such as council minutes that are not yet widely available on the internet to user U. The learning model 109 may also be customized by the extended generation unit 110 or input prompts.
[0069] Although the example shows the information provision server 10 having multiple learning models 109, the information provision server 10 only needs to have at least one learning model 109. Furthermore, the information provision server 10 may also have multiple extension generation units 110, each corresponding to one of the multiple learning models 109. This allows for customization of each type of learning model 109.
[0070] The learning model 109 may be machine-learned based on the training data stored in the database 111. In this way, newly accumulated information in the database 111 will be included in the learning model 109, and the learning model 109 will be able to generate answers based on the latest information. Alternatively, newly accumulated information in the database 111 may be incorporated into the extended generation unit 110.
[0071] Database 111 stores the information necessary for the learning model 109 to generate answers. The information stored in this database 111 includes publicly available information on the internet, identity verification media 41, and information obtainable from the certificate verification server 20 or the resident information management platform 50 (attribute information, non-public information, and unique information). For example, when the information provision server 10 assigns a user-specific ID to user U, it obtains information related to user U from the resident information management platform 50 and registers it in database 111. In this way, the learning model 109 can immediately generate answers based on information unique to user U.
[0072] Figure 4 shows the contents registered in database 111. The User-Specific ID field registers a unique user ID, which is a single identifier that can individually identify each user U and remains unchanged throughout their life. This User-Specific ID serves as the primary key for the tables registered in database 111. Each user U is assigned a User-Specific ID. Various information is registered using this User-Specific ID as the primary key.
[0073] For example, information such as user ID, identity verification information, available models, available amount (upper limit), and current usage (current value) is registered in database 111. Other information, such as user U's attribute information, may also be registered in database 111.
[0074] The identity verification information section will register the four basic pieces of information necessary for verifying User U's identity (name, address, date of birth, and gender). Note that the identity verification information may also include an email address and other information.
[0075] The "Available Models" section registers the types of learning models 109 that user U can use. The types of learning models 109 registered in this section are determined, for example, based on user U's attribute information. In this way, the most suitable type of learning model 109 can be used for each individual user U.
[0076] The "Available Amount (Upper Limit)" field registers the available amount, or upper limit, for each type of learning model 109. This available amount is the upper limit of the number of times user U can use the learning model 109 within a predetermined period. In other words, it is the upper limit of the number of times user U can make requests (questions) to the learning model 109. This upper limit is predetermined based on attribute information indicating user U's attributes. Furthermore, the upper limit may be changed as appropriate in response to changes in user U's attribute information.
[0077] The "Current Usage (Current Value)" field registers the current usage and value for each type of learning model 109. This current value is reset after a certain period of time or after certain conditions are met. For example, the current value may be reset when the date changes, or after a predetermined period of time has elapsed since the last time user U used it. Alternatively, the current value may be reset when it is proven that the user has become a resident of the municipality due to relocation, etc., or when a predetermined additional fee is paid.
[0078] For example, the available amount (upper limit) and usage amount (current value) are set for a predetermined type of first learning model. The available amount (upper limit) and usage amount (current value) are set for a second learning model of a different type from this first learning model.
[0079] Furthermore, the available quantity of all types of learning models 109 may be managed by a single item (a single upper limit). Alternatively, the usage of all types of learning models 109 may be managed by a single item (a single current value).
[0080] Furthermore, the information provision system 1 may use other methods to implement the function of controlling usage. For example, instead of providing the available amount (upper limit) and usage amount (current value), the information provision system 1 may grant the user U a usage coupon with an expiration date after a certain period of time has elapsed or certain conditions have been met.
[0081] The specified organization in this embodiment includes at least one of an administrative agency G or a private organization. In this way, output information including at least one of administrative services, private services, and local information can be provided to user U.
[0082] The output information includes non-public information held by a specific institution. In this way, it is possible to construct a learning model 109 that is tuned specifically for that institution.
[0083] Furthermore, the output information includes information specific to user U that is held by a particular institution. In this way, customized output information can be provided according to user U. This specific information includes, for example, information on health checkups, benefits, pensions, and tax payments.
[0084] Furthermore, the output information is generated by the learning model 109 after the response has been generated including information specific to user U. In this way, customized output information can be provided according to user U. This specific information includes, for example, the weather in user U's place of residence, health advice tailored to user U's age group, and tourist information.
[0085] (Example hardware configuration) The information provision server 10, which is a computer, may have the configuration shown in Figure 3. The information provision server 10 has a bus 1010, a processor 1020, memory 1030, a storage device 1040, an input / output interface 1050, and a network interface 1060.
[0086] Bus 1010 is a data transmission path for the processor 1020, memory 1030, storage device 1040, input / output interface 1050, and network interface 1060 to send and receive data to and from each other. However, the method of connecting the processor 1020 and the other components to each other is not limited to bus connection.
[0087] Processor 1020 is a circuit that includes arithmetic units such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit).
[0088] Memory 1030 is a main memory device implemented using RAM (Random Access Memory), etc.
[0089] The storage device 1040 is a removable media such as an HDD (Hard Disk Drive), SSD (Solid State Drive), flash memory, or memory card, or an auxiliary storage device such as ROM (Read Only Memory), and has a recording medium. The recording medium of the storage device 1040 stores programs that implement each function of the information provision server 10 (such as the user authentication unit 101).
[0090] The processor 1020 reads this program into memory 1030 and executes it. This causes the processor 1020 to perform the function corresponding to this program. In other words, the program stored in memory 1030 causes the information provision server 10 to perform a predetermined function.
[0091] The input / output interface 1050 connects the information provision server 10 to a predetermined input / output device. The input / output device is, for example, an input device such as a keyboard, an output device such as a display, or an input / output device in which a touch panel is superimposed on a display.
[0092] The network interface 1060 is an interface for connecting the information provision server 10 to a predetermined communication network. This communication network may be, for example, the Internet, a LAN (Local Area Network), or a WAN (Wide Area Network). The method by which the network interface 1060 connects to the communication network may be wireless or wired.
[0093] The information provision server 10 has been described above. In addition to the above configuration, the information provision server 10 may have an information input device for user U to input various information into the information provision server 10 through operation. The information input device may be, for example, a keyboard, mouse, or touch panel. The information provision server 10 may also have a display, speaker, vibration motor, or LED (light-emitting diode) for showing various information to user U.
[0094] The configuration of the information provision system 1 does not necessarily have to be provided on a single information provision server 10. For example, one information provision system 1 may be implemented using multiple computers connected to each other via a network N. Alternatively, other computers connected to the information provision server 10 via network N may also be equipped with a learning model 109.
[0095] Some components of the information provision system 1 may be provided on the administrator terminal 30 or the user terminal 40.
[0096] (Information processing) Next, the information provision process performed by the information provision server 10 will be explained using the flowchart in Figure 5. The aforementioned diagrams will be used as references as appropriate. The following steps represent at least some of the processes included in the information provision process; other steps may also be included.
[0097] First, in step S1, when the user terminal 40 accesses the user's terminal (request for identity verification), the user U's identity is verified (identity verification authentication) using the public personal authentication service via the identity verification medium 41 (identity verification process).
[0098] In the next step S2, the identification information assignment unit 102 of the information provision server 10 obtains attribute information indicating the attributes of user U from the identity verification medium 41, a system managed by a public institution, or a database 111 (identification information assignment process). The identification information assignment unit 102 may also obtain attribute information stored in the identity verification medium 41 through authentication by the identity verification medium 41. Alternatively, the identification information assignment unit 102 may obtain attribute information that has been obtained from the identity verification medium 41 and then stored in the database 111.
[0099] In the next step S3, the identification information assignment unit 102 of the information provision server 10 refers to the history of changes to the serial numbers of electronic certificates, or the relationships between the serial numbers of multiple types of electronic certificates, using the serial number linking function provided by J-LIS (identification information assignment process).
[0100] In the next step S4, the identification information assignment unit 102 of the information provision server 10 determines, based on the serial number linking function described above, whether or not a user-specific ID has already been generated, that is, whether or not user U is the same person who has previously registered with the information provision server 10 (identification information assignment process). If a user-specific ID has already been generated, that is, if user U is the same person who has previously registered with the information provision server 10 (YES in step S4), the process proceeds to step S8. On the other hand, if a user-specific ID has not already been generated, that is, if user U is not the same person who has previously registered with the information provision server 10 (NO in step S4), the process proceeds to step S5.
[0101] In step S8, which proceeds if the answer in step S4 is YES, the identification information assignment unit 102 of the information provision server 10 returns the generated user-specific ID to user U and proceeds to step S7.
[0102] In step S5, which proceeds if the answer in step S4 is NO, the identification information assignment unit 102 of the information provision server 10 newly assigns a user-specific ID as identification information to user U (identification information assignment process).
[0103] In the next step S6, the identification information assignment unit 102 of the information provision server 10 sets an available amount (upper limit) corresponding to the user-specific ID assigned to user U, and registers this available amount in the database 111 along with other information. At this time, the identification information assignment unit 102 may increase or decrease this available amount based on user U's attribute information (address, date of birth, age, place of residence, tax domicile, place of origin, nationality, etc.).
[0104] In the next step S7, the information provision unit 108 of the information provision server 10 executes the learning model usage process and completes the information provision process.
[0105] (Process using the learned model) Next, the learning model usage process executed by the information provision server 10 will be explained using the flowchart in Figure 6. The aforementioned diagrams will be used as references as appropriate. The following steps represent at least some of the processes included in the learning model usage process; other steps may also be included.
[0106] First, in step S11, the information provision unit 108 of the information provision server 10 receives a request (question) input from user U via user terminal 40.
[0107] In the next step S12, the model switching unit 103 of the information provision server 10 switches the type of learning model 109 to enable the use of the optimal learning model 109 based on the attribute information of user U. Alternatively, the model switching unit 103 may accept the type of learning model 109 that user U wishes to use and switch to that type of learning model 109. In other words, user U may be allowed to select the type of learning model 109 as appropriate.
[0108] In the next step S13, the usage control unit 105 of the information provision server 10 refers to the database 111 and determines whether the usage (current value) of the learning model 109 by user U exceeds the available amount (upper limit) (usage control process). If it exceeds the available amount (upper limit) (if YES in step S13), the process proceeds to step S18. On the other hand, if it does not exceed the available amount (upper limit) (if NO in step S13), the process proceeds to step S14.
[0109] In step S18, which proceeds if the answer in step S13 is YES, the restriction release unit 106 of the information provision server 10 determines whether or not the usage limit is currently being released (restriction release process). If the usage limit is currently being released (if the answer in step S18 is YES), the process proceeds to step S14. On the other hand, if the usage limit is not currently being released (if the answer in step S18 is NO), the process proceeds to step S19.
[0110] In step S19, which proceeds if the answer in step S18 is NO, the restriction release unit 106 of the information provision server 10 notifies user U that the usage limit has been reached. For example, the restriction release unit 106 displays a message on the user terminal 40 indicating that the usage limit has been reached and completes the learning model usage process.
[0111] In step S14, which proceeds if the answer in step S13 is NO or if the answer in step S18 is YES, the information provision unit 108 of the information provision server 10 generates an answer (output information) using the learning model 109 (information provision processing).
[0112] In the next step S15, the usage control unit 105 of the information provision server 10 adds "1" to the number of times the learning model 109 has been used by user U (current usage value) registered in the database 111 (usage control process).
[0113] In the next step S16, the information provision management unit 107 of the information provision server 10 determines whether the answer (output information) generated by the learning model 109 is an answer that requires confirmation by administrator M. If the answer is an answer that requires confirmation by administrator M (YES in step S16), the process proceeds to step S20. On the other hand, if the answer is not an answer that requires confirmation by administrator M (NO in step S16), the process proceeds to step S17.
[0114] In step S17, which proceeds if the answer in step S16 is NO, the information provision unit 108 of the information provision server 10 provides the user U with the answer (output information) generated by the learning model 109. For example, the information provision unit 108 displays the answer on the display of the user terminal 40 and completes the learning model usage process.
[0115] In step S20, which proceeds if the answer in step S16 is YES, the information provision management unit 107 of the information provision server 10 notifies user U that it will provide a response (output information) at a later date. For example, the information provision management unit 107 displays a message on the user terminal 40 indicating that it will provide a response at a later date.
[0116] In the next step S21, the information provision management unit 107 of the information provision server 10 makes an approval decision or edits the provision of the answer (output information) generated by the learning model 109 based on the operation of administrator M. Administrator M uses the administrator terminal 30 to make an approval decision or edit the provision of the answer generated by the learning model 109.
[0117] In the next step S22, the information provision server 10's provision management unit 107 sets the response to be provided to user U at a later date after administrator M has completed the approval decision or editing of the response (output information) and completes the learning model usage process. When user U accesses the information provision server 10 at a later date, they can receive the pre-configured response.
[0118] As described above, according to this embodiment, by controlling the amount of usage according to user U, fairness in the use of the learning model 109 can be ensured.
[0119] In the above-described embodiment, the identification information assignment unit 102 assigns one unique user ID to a single user U that remains unchanged throughout their life, but other configurations are also possible. For example, the identification information assignment unit 102 may assign two or more unique user IDs to a single user U. Furthermore, the unique user ID may be changed when certain conditions are met.
[0120] While the flowchart above illustrates a configuration where each step is executed sequentially, the order of each step is not necessarily fixed, and the order of some steps may be reversed. Furthermore, some steps may be executed in parallel with others. Also, the steps included in the flowchart above represent at least a subset of the steps, and other steps may be included in the flowchart.
[0121] The flowchart described above represents a process that is repeated at regular intervals. This repeated process executes the information provision method in Information Provision System 1. Note that this process may be interrupted and executed while Information Provision System 1 is running other main processes.
[0122] The aforementioned system may include a computer equipped with artificial intelligence (AI) for machine learning. Furthermore, the aforementioned system may include a deep learning unit that extracts specific patterns from multiple patterns based on deep learning.
[0123] The aforementioned computer-based analysis can utilize analytical techniques based on artificial intelligence learning. For example, trained models generated by machine learning using neural networks, trained models generated by other machine learning methods, deep learning algorithms, and mathematical algorithms such as regression analysis can be used. Furthermore, forms of machine learning include clustering and deep learning.
[0124] The aforementioned system includes a computer equipped with artificial intelligence that performs machine learning. For example, this system may consist of one computer equipped with a neural network, or it may consist of multiple computers equipped with neural networks.
[0125] Here, a neural network is a mathematical model that represents the characteristics of brain function through computer simulation. For example, it shows a model in which artificial neurons (nodes) that form a network through synaptic connections change the strength of their synaptic connections through learning and acquire problem-solving abilities. Furthermore, neural networks acquire problem-solving abilities through deep learning.
[0126] For example, a neural network may have multiple layers, each consisting of several units. By pre-training a multi-layer neural network with training data (supervised data), it is possible to automatically extract features from patterns of changes in the state of a circuit or system. Furthermore, the number of hidden layers, units, learning rate, number of training iterations, and activation function of a multi-layer neural network can be set arbitrarily via the user interface.
[0127] Furthermore, a deep reinforcement learning approach may be used in the neural network, in which a reward function is set for each information item to be learned, and the information item with the highest value is extracted based on the reward function.
[0128] Furthermore, there are various machine learning techniques, such as autoencoders, LSTM (Long Short-Term Memory), SDF (Signed Distance Function), GAN (Generative Adversarial Network), and RNN (Recurrent Neural Network). These techniques may be applied to the machine learning in this embodiment.
[0129] The learning model 109 includes an input layer, a hidden layer, and an output layer. Input data is input to the input layer. The parameters of the hidden layer are pre-machine-trained using the training data. The output layer outputs output data that shows the results processed by the hidden layer in response to the input data input to the input layer.
[0130] Each of the aforementioned components of the system may be implemented with dedicated hardware. Furthermore, some or all of each component may be implemented by general-purpose or dedicated circuits, processors, or combinations thereof. These may be comprised of a single chip or multiple chips connected via a bus. Some or all of each component may be implemented by a combination of the aforementioned circuits and programs. Additionally, a CPU (Central Processing Unit), GPU (Graphics Processing Unit), etc., can be used as the processor. Moreover, at least some of the functions of this embodiment may be provided in the form of IaaS (Infrastructure as a Service), PaaS (Platform as a Service), or SaaS (Software as a Service).
[0131] The aforementioned program, when loaded into a computer, includes a set of instructions (or software code) for causing the computer to perform one or more of the functions described in the embodiments. The program may be stored on a non-temporary computer-readable medium or a physical storage medium. Examples, but not limited to, include computer-readable mediums or physical storage mediums such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, SSD (solid-state drive), or other memory technologies. Examples, but not limited to, include CD-ROMs, DVDs (digital versatile discs), Blu-ray discs, or other optical disc storage. Examples, but not limited to, include magnetic tape, magnetic disk storage, or other magnetic storage devices. The program may be transmitted over a temporary computer-readable medium or a communication medium. Examples, but not limited to, include temporary computer-readable mediums or communication mediums such as electrical, optical, acoustic, or other forms of propagating signals.
[0132] The embodiments have been described above, but the configurations of the embodiments described above may be combined with each other, or some of the configurations may be replaced with other configurations. Furthermore, the configurations of the embodiments described above may be modified in various ways without departing from the spirit of the invention.
[0133] Each drawing is merely illustrative to illustrate one or more embodiments. Each drawing may be associated with one or more other embodiments rather than with only one specific embodiment. As those skilled in the art will understand, various features or steps described with reference to any one drawing can be combined with features or steps shown in one or more other drawings to create embodiments not explicitly illustrated or described. Not all features or steps shown in any one drawing to illustrate an exemplary embodiment are necessarily required, and some features or steps may be omitted. The order of steps shown in any of the drawings may be changed as appropriate. [Explanation of Symbols]
[0134] 1. Information Provision System 10 Information Provision Server 20 Certificate Verification Servers 30 Administrator terminals 40 User terminals 41 Identity verification medium 50 Resident Information Management Platform 101 Identity Verification Department 102 Identification Information Assignment Unit 103 Model switching section 104 Audio and Video Processing Unit 105 Usage Control Unit 106 Restriction Release Section 107 Provision Management Department 108 Information Provision Department 109 Learning Models 110 Extended generation unit 111 Database 1010 Bus 1020 processor 1030 memory 1040 Storage Devices 1050 Input / Output Interface 1060 Network Interfaces G Government agencies M Administrator N Network U User
Claims
1. An information provision system in which a specific organization provides information to external users, A user authentication unit that authenticates the user's identity using a medium capable of verifying identity, An identification information assignment unit that assigns identification information to the user based on the aforementioned authentication, A usage control unit that controls the usage of a machine learning model that provides the user with output information output in response to the user's request input based on the aforementioned identification information, An information provision system equipped with the following features.
2. In the information provision system described in claim 1, The aforementioned identity verification unit performs identity verification of the user using the public personal authentication service provided by the identity verification medium. The information provision system provides the identification information assignment unit, when assigning the identification information to the same user whose information on the identity verification medium has been updated, based on the identity verification medium, assigns the same identification information as the previously assigned identification information.
3. In the information provision system according to claim 1 or claim 2, The usage control unit is an information provision system that increases or decreases the usage amount based on attribute information indicating the user's attributes, which can be obtained from the personal identification medium or a system managed by a public institution.
4. In the information provision system described in claim 3, The aforementioned learning model consists of multiple types of the aforementioned learning models, An information provision system comprising a model switching unit that switches the type of learning model based on the attribute information and outputs the output information.
5. In the information provision system according to claim 1 or claim 2, The output information is provided by an information provision system that includes non-public information held by the specified organization.
6. In the information provision system according to claim 1 or claim 2, The output information includes information specific to the user held by the designated organization, and is provided by an information provision system.
7. In the information provision system according to claim 1 or claim 2, The aforementioned output information is output from the learning model as a result of answer generation including information specific to the user, in an information provision system.
8. In the information provision system according to claim 1 or claim 2, The learning model is an information provision system that, in response to input requests from the user via at least one of voice or video, outputs output information via at least one of voice or video.
9. In the information provision system according to claim 1 or claim 2, An information provision system comprising a restriction release unit that releases the usage limit based on at least one of predetermined conditions or operations performed by the specified organization.
10. In the information provision system according to claim 1 or claim 2, The output information can be reviewed by an internal administrator of the specified organization before it is provided to the user. An information provision system comprising a provision management unit that allows the administrator to make approval decisions regarding provision or to edit based on the output information.
11. In the information provision system according to claim 1 or claim 2, The aforementioned specified organization is an information provision system that includes at least one of a government agency or a private organization.
12. One or more computers that implement an information provision system in which a specific organization provides information to external users, An identity verification process that authenticates the user's identity using a media capable of verifying identity, Based on the aforementioned authentication of the user, an identification information assignment process is performed to assign identification information to the user, A usage control process that controls the usage of a machine learning model that provides the user with output information output in response to the user's request input, based on the aforementioned identification information. A method of providing information to carry out the task.
13. One or more computers that implement an information provision system in which a specific organization provides information to external users, An identity verification process that authenticates the user's identity using a media capable of verifying identity, Based on the aforementioned authentication of the user, an identification information assignment process is performed to assign identification information to the user, A usage control process that controls the usage of a machine learning model that provides the user with output information output in response to the user's request input, based on the aforementioned identification information. An information-providing program that initiates the execution of [something].