Display control system, control method for display control system, and program

Abstract

This makes it easier to motivate data owners to allow others to use their data for machine learning. [Solution] The display control unit 121a displays information on the display unit 125a indicating permission to allow another user (second user) to use the user's (first user's) training data for training the learning model. At this time, the display control system displays the information indicating permission and the benefit information indicating the benefits of permission on the display unit 125a in a corresponding manner.

Description

[Technical Field]

[0001] The present disclosure relates to a display control system, a control method for a display control system, and a program. [Background Art]

[0002] As a technique for utilizing data used for machine learning, there is the technique described in Patent Document 1. Patent Document 1 discloses that a user of an imaging device permits others to use a learning source image (test data) and specifies charging conditions for others' use of the learning source image by operating the imaging device. [Prior Art Literature] [Patent Literature]

[0003] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2020-166744 [Summary of the Invention] [Problem to be Solved by the Invention]

[0004] However, Patent Document 1 does not specifically describe what information is to be displayed and what operation is to be performed when the owner of the learning source image performs an operation for permitting others to use the learning source image. Further, in the technique described in Patent Document 1, the consideration for permitting others to use the learning source image is limited to charging, and the owner of the learning source image himself / herself specifies the charging conditions. Therefore, there is a problem that it is difficult to provide motivation to the owner of data used for machine learning to allow others to use the data.

[0005] The present disclosure has been made in view of the above problems, and an object of the present disclosure is to make it easy to provide motivation to the owner of data used for machine learning to allow others to use the data. [Means for Solving the Problem]

[0006] A first example of a display control system of the present disclosure is a display control system comprising a reception unit for receiving first training data of a first user, and a display control means for displaying permission information on a display unit indicating permission to use the first training data for training a training model of a second user different from the first user, wherein the display control means displays the permission information and benefit information indicating the benefits of granting permission in correspondence with each other on the display unit. A second example of the display control system of the present disclosure is a display control system comprising a reception unit for receiving first training data of a first user, and display control means for displaying disallowance information on a display unit indicating that the first training data is not permitted to be used for training a training model of a second user different from the first user, wherein the display control means displays the disallowance information and disadvantage information indicating the disadvantages of the disallowance on the display unit in correspondence with each other.

[0007] A first example of a control method for a display control system of the present disclosure is a control method for a display control system comprising a receiving unit for receiving first learning data of a first user, comprising a display control step of displaying permission information on a display unit indicating permission to use the first learning data for learning a learning model of a second user different from the first user, wherein in the display control step, the permission information and benefit information indicating the benefits of granting permission are displayed on the display unit in a corresponding manner. A second example of a control method for a display control system of the present disclosure is a control method for a display control system comprising a receiving unit for receiving first learning data of a first user, comprising a display control step of displaying disallowance information on a display unit indicating that the use of the first learning data for learning a learning model of a second user different from the first user is not permitted, characterized in that in the display control step, the disallowance information and disadvantage information indicating the disadvantages of the disallowance are displayed on the display unit in a corresponding manner.

[0008] The program of this disclosure causes a computer to function as a display control means for the display control system. [Effects of the Invention]

[0009] According to this disclosure, it is possible to make it easier to motivate the owner of the data to allow others to use the data used for machine learning. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example of the configuration of a display control system. [Figure 2] This figure shows an example of the functional configuration of a display control system. [Figure 3] This figure shows an example of a set of information related to a learning model. [Figure 4] This figure shows an example of a set of information related to training data. [Figure 5A] This figure shows an example of the initial permission settings screen. [Figure 5B] This figure shows an example of the usage permission settings screen after the input has been completed. [Figure 6] This figure shows an example of a screen for viewing training data. [Figure 7] This is a flowchart illustrating an example of a control method for a display control system. [Modes for carrying out the invention]

[0011] Hereinafter, an embodiment of this disclosure will be described with reference to the drawings. Figure 1 shows an example of the configuration of the display control system according to this embodiment. Figure 1 illustrates a case where the display control system comprises a learning model provider device 110, learning model user devices 120a to 120b, and a language model server 130. In Figure 1, etc., "..." is a symbol indicating that similar devices may exist above, below, or to the left or right of it. In Figure 1, "..." indicates that there may be other model user devices separate from the learning model user devices 120a to 120b. The number of learning model user devices 120a to 120b is not limited to two; it may be one or three or more. In this embodiment, for the sake of simplicity, we illustrate a case where there is one user for one learning model user device. However, this is not necessarily required. For example, multiple users may log in to the same learning model user device with their own accounts. In this case, the learning model user device may process and store information for each of the multiple users. This may be done if the display control system has one learning model user device, or if the display control system has two or more learning model user devices, this may be done in at least one of those two or more learning model user devices.

[0012] Furthermore, Figure 1 illustrates a case where the learning model provider device 110, the learning model user devices 120a to 120b, and the language model server 130 are connected to each other via a network 140 including the Internet. The learning model provider device 110 and the language model server 130 may reside on the cloud. Note that the configuration of the display control system is not limited to the configuration shown in Figure 1. For example, the language model server 130 may be included in the learning model provider device 110. Also, for example, one of the learning model user devices 120a to 120b may be included in the learning model provider device 110.

[0013] The learning model provider device 110 creates and provides a learning model to the learning model user devices 120a to 120b. Providing a learning model means, for example, making it possible for users of the learning model user devices 120a to 120b to know the results of inference using the learning model. For example, the learning model provider device 110 may download the learning model to the learning model user devices 120a to 120b, or it may make the learning model user devices 120a to 120b able to use the learning model online. In the latter case, for example, the learning model provider device 110 may calculate the dependent variable data by using the learning model based on the independent variable data provided by the learning model user devices 120a to 120b, and send the dependent variable data to the learning model as the result of inference.

[0014] The learning model is a so-called pre-trained model, created, for example, by performing machine learning on training data to determine the relationship between independent variables (explanatory variables) and dependent variables (target variables). The independent and dependent variables may be quantitative variables (variables expressed numerically) or qualitative variables (variables expressed as information other than numerical values). In this embodiment, the case where the machine learning method is supervised learning is given as an example. However, the machine learning method is not limited to supervised learning; for example, it may be unsupervised learning or reinforcement learning. Furthermore, the type of learning model is not limited; for example, the learning model provider device 110 can provide various learning models such as deep learning learning models, regression analysis learning models, decision trees, k-means algorithm learning models, SVM (Support Vector Machine) learning models, k-nearest neighbors learning models, Q-learning learning models, etc.

[0015] Furthermore, the objects to be derived (estimated) by the learning model are not limited. For example, the learning model may be one that estimates the state of an object such as a device (e.g., anomaly), or it may be a model that estimates objects such as people contained in an image such as a captured image. Estimation may involve deriving the current value of the dependent variable, or it may involve deriving (i.e., predicting) the future value of the dependent variable.

[0016] The learning model user-side device 120a provides training data prepared by a user of the learning model user-side device 120a to the learning model provider-side device 110.

[0017] The training data only needs to be data used for training the learning model. In the present embodiment, a case where the training data is learning data to be trained by the learning model is exemplified. In this case, in addition to the learning data (data to be trained), the training data may include at least one of data for adjusting hyperparameters and data for evaluating a learning result (performance of the learning model).

[0018] Further, as described above, in the present embodiment, a case where the learning model is created by supervised learning is exemplified. Therefore, in the present embodiment, a case where a correct label is included in the training data (learning data) (i.e., a case where annotation has been performed) is exemplified. However, this is not necessarily required. For example, the learning model provider-side device 110 may perform annotation. Note that when the learning model performs unsupervised learning, no correct label is required. Further, the training data is not limited to learning data to be trained by the learning model as long as it is data used for training the learning model (at the time of training). For example, the training data may be a hyperparameter.

[0019] Note that in the present embodiment, a set of data of the number required for one training session is referred to as training data. As described above, in the present embodiment, a case where the learning model is created by supervised learning and a correct label is included in the training data (learning data) is exemplified. In this case, the training data (learning data) includes the number of pairs required for one training session, each pair consisting of training target data (information indicating content of an independent variable) and a correct label (information indicating content of a dependent variable).

[0020] In this embodiment, we illustrate a case where the learning model user device 120b, other than the learning model user device 120a, also has the same functions as the learning model user device 120a. The learning model user device 120b provides the learning model provider device 110 with the learning data prepared by the user of the learning model user device 120b.

[0021] In the following explanation, from the perspective of the users of the learning model-using devices 120a and 120b, the learning model created (newly) using the training data they have prepared will be referred to as "their own learning model," as necessary. Furthermore, in the following explanation, from the perspective of the users of the learning model-using devices 120a and 120b, the training data used when creating the new learning model (their own learning model) will be referred to as "their own training data," as necessary. In this case, "their own training data" is an example of the first training data.

[0022] Furthermore, in the following explanation, from the perspective of a user of learning model-using device 120b other than learning model-using devices 120a and 120b, the user will refer to themselves as "their own user" as appropriate, and other users as "other users" as appropriate. In this case, the "own user" is an example of the first user, and another user is an example of the second user.

[0023] Furthermore, in the following explanation, training data prepared by other users will be referred to as "other users' training data" as needed, from the perspective of the user. In this case, "other users' training data" is an example of second training data. Furthermore, in the following explanation, a training model created (newly) using the training data prepared by another user (other users' training data) will be referred to as "other users' training model" as needed, from the perspective of the user.

[0024] As described above, in the following explanation, the same user may be referred to as "my user" or "another user." Similarly, the same training data may be referred to as "my own training data" or "the training data of another person or another user." Furthermore, the same learning model may be referred to as "my own training model" or "the training model of another person or another user."

[0025] In this embodiment, we illustrate a case in which the learning model provider device 110 uses the training data provided by the learning model user devices 120a and 120b to create (new) learning models for the users of the learning model user devices 120a and 120b, respectively.

[0026] Furthermore, this embodiment illustrates a case where the learning model provider device 110, after creating a learning model for a certain user (the user), can update (retrain) the user's own learning model using learning data prepared by another user (another user's learning data). However, this embodiment illustrates a case where such updating of the learning model is limited to cases where the other user has given permission.

[0027] Furthermore, this embodiment illustrates a case in which the learning model providing device 110 can provide a certain user (the user) with another user's learning model (another person's learning model). However, in this embodiment, the provision of such another user's learning model is limited to cases where the other user has given permission.

[0028] The value of the display control system can be increased if each user is encouraged to allow other users to use their own training data. In this case, for example, the possibility of an increase in the number of users participating in the display control system can be increased. Therefore, in this embodiment, we illustrate a case in which users of the learning model user devices 120a to 120b display mutually corresponding information on the display units 125a to 125b of the learning model user devices 120a to 120b, showing the benefits of allowing other users to use their own training data and permission information indicating permission to use their own training data. In addition, in this embodiment, we illustrate a case in which users of the learning model user devices 120a to 120b display mutually corresponding information on the display units 125a to 125b of the learning model user devices 120a to 120b, showing the disadvantages of not allowing other users to use their own training data and denial information indicating that permission to use their own training data is not permitted. While displaying both the advantages and disadvantages is preferable as it encourages users to share their learning data with other users, displaying only one of them is also acceptable.

[0029] In this embodiment, we illustrate a case where information including merit information, permission information, disadvantage information, and denial information, which serves as a decision-making criterion for granting permission to other users to use one's own training data, is displayed within a single frame (window). In the following description, this decision-making criterion information will be referred to as decision-making criterion information as needed. Also, in the following description, this frame will be referred to as the usage permission setting screen as needed.

[0030] In Figure 1, the language model server 130, in response to requests from external devices such as the learning model provider device 110, causes a language model trained by machine learning using a large amount of data to perform natural language processing, and transmits the results of the processing to the external device. By using the language model, for example, machine translation, text generation, and answering questions can be performed automatically. The language model 131 used in this embodiment is a trained model, such as a large language model (LLM) like GPT4-o(registered trademark), LaMDA(registered trademark), LLaMA(registered trademark), and LLaVA(registered trademark). In this embodiment, we illustrate a case in which the language model server 130 derives at least one of the judgment indicators to be displayed on the usage permission setting screen by using the language model. As will be described in detail later, more specifically in this embodiment, we illustrate a case in which the language model server 130 derives the degree of the degree of the degree of the price reduction of its own learning model.

[0031] Figure 2 shows an example of the functional configuration of a display control system. In Figure 2, this embodiment illustrates a case where the learning model providing device 110 comprises a processing unit 111, a receiving unit 112, a providing unit 113, and a storage unit 114. This embodiment also illustrates a case where the learning model consuming device 120a comprises a processing unit 121, a receiving unit 122, a providing unit 123, and a storage unit 124. As mentioned above, the learning model consuming devices 120a to 120b can be implemented with devices having the same configuration. Therefore, in this embodiment, a detailed explanation of the configuration of the learning model consuming device 120b other than the learning model consuming device 120a is omitted. Furthermore, this embodiment illustrates a case where the language model server 130 comprises a language model 131.

[0032] The learning model provider device 110 includes, for example, one or more hardware processors such as a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit), and one or more memory modules such as RAM (Random Access Memory) and ROM (Read Only Memory). It performs various calculations by executing one or more programs stored in memory using one or more hardware processors. The learning model provider device 110 also includes, for example, a NIC (Network Interface Card) for connecting the learning model provider device 110 to a network.

[0033] Furthermore, the learning model providing device 110 may be connected to a storage medium that can be read by the learning model providing device 110 (computer). Note that this storage medium is not limited to a storage medium that can only read data; it may also be a storage medium that can be read and written to by the learning model providing device 110 (computer). Additionally, the learning model providing device 110 may be connected to an output device such as a computer display. Furthermore, the learning model providing device 110 may be connected to an input device such as a keyboard or mouse. Information input to the learning model providing device 110 may also be performed using a GUI (Graphical User Interface).

[0034] Furthermore, the learning model providing device 110 may be implemented using dedicated hardware such as an ASIC (Application Specific Integrated Circuit) and an FPGA (Field Programmable Gate Array). Furthermore, the learning model user devices 120a to 120b and the language model server 130 can also be implemented, for example, by using the same hardware as the learning model provider device 110.

[0035] [Learning model provider device 110] The reception unit 112 receives various types of information from external devices. As shown in Figure 1, this embodiment illustrates a case where the reception unit 112 receives information transmitted from an external device via the network 140. This embodiment also illustrates a case where the reception unit 112 functions as a user interface for the learning model providing device 110. In this case, the reception unit 112 receives input operations from the user of the learning model providing device 110.

[0036] The information provision unit 113 provides various types of information to an external device. As shown in Figure 1, this embodiment illustrates a case where the information provision unit 113 transmits information to an external device via the network 140. The processing unit 111 provides overall control of the learning model providing device 110. In this embodiment, an example is given in which the processing unit 111 performs the creation and management of learning models, the management of learning data, and the creation of information necessary for displaying the usage permission setting screen.

[0037] The storage unit 114 stores various types of information necessary for processing by the learning model providing device 110. In this embodiment, an example is given in which the storage unit 114 includes a learning model storage unit 114a, a learning model-related information storage unit 114b, a learning data storage unit 114c, and a learning data-related information storage unit 114d.

[0038] The learning model storage unit 114a stores, for example, the user IDs (identification information) of the learning model user devices 120a to 120b, the IDs of the learning models of those users, and the main information of those learning models, each associated with each other. The main information of the learning model refers to the information that constitutes the learning model, and for example, the information necessary to calculate the dependent variables, excluding the independent variables. In this embodiment, an example is given in which the ID of the platform provider of the display control system (the user of the learning model provider device 110), the ID of the learning model of the platform provider, and the main information of the learning model of the platform provider are also stored in the learning model storage unit 114a.

[0039] The learning model-related information storage unit 114b stores learning model-related information for each learning model whose main information is stored in the learning model storage unit 114a. Learning model-related information is information managed by the learning model provider device 110 as information related to the learning model, separate from the main information of the learning model. In this embodiment, an example is given in which the learning model-related information includes information necessary to display the usage permission setting screen described later.

[0040] Figure 3 shows an example of the learning model-related information group 300. In Figure 3 and other figures, "***" is a symbol indicating that specific information (numbers, letters, etc.) has been omitted. Also, as mentioned above, "···" is a symbol indicating that similar information may exist above and below it.

[0041] Figure 3 illustrates a case where the learning model-related information 310a to 310b includes user ID 301, learning model ID 302, content 303, profile 304, model user 305, data provider 306, benefit 307, and evaluation 308.

[0042] User ID 301 is the user ID of the learning model. In this embodiment, an example is given in which a user who wishes to create a learning model (any one of the learning model user devices 120a to 120b) provides training data to be used for training the learning model, and transmits the user's ID and name, and the name of the training data, to the learning model provider device 110. In this case, for example, the receiving unit 112 receives the user's ID, and the processing unit 111 stores the ID in the user ID 301 field (storage area).

[0043] The learning model ID 302 is the ID of the learning model. In this embodiment, we illustrate the case in which the processing unit 111 assigns an ID to a learning model when it creates a learning model using training data provided by a user (any one of the learning model user devices 120a to 120b) who wishes to create a learning model. In this case, the processing unit 111 stores the ID assigned to the learning model in the field of learning model ID 302 that is in the same row (column) as the field of user ID 301 in which the user's ID is stored. In this case, the processing unit 111 may also notify the user of the learning model (any one of the learning model user devices 120a to 120b) of the ID of the user's learning model via the provision unit 113.

[0044] Content 303 is information that concisely describes the learning model. For example, if the learning model is a model that estimates equipment abnormalities, the string "Abnormality Estimation Model" is stored as content. In this embodiment, an example is given in which a user who wishes to create a learning model (any one of the learning model user devices 120a to 120b) sends information indicating the content to the learning model provider device 110 when providing training data to be used for training the learning model. In this case, the receiving unit 112 receives the information indicating the content, and the processing unit 111 stores the information in the field of content 303 that is on the same row as the field of learning model ID 302 in which the ID of the learning model is stored.

[0045] Profile 304 is information that provides a more detailed overview of the learning model than the content. For example, if the learning model is a model that estimates abnormalities in a heat exchanger, the string "Heat exchanger abnormality estimation model" may be stored as Profile 304. Also, if the learning model is a model that estimates abnormalities in the heat exchanger fan, the string "Heat exchanger fan abnormality estimation model" may be stored as Profile 304. In this embodiment, an example is given in which a user who wishes to create a learning model (any one of the learning model user devices 120a to 120b) sends information indicating the contents of the profile (information that provides a more detailed overview of the learning model than the content) to the learning model provider device 110 when providing training data to be used for training the learning model. In this case, for example, the receiving unit 112 receives the information indicating the contents of the profile, and the processing unit 111 stores the information in a field of Profile 304 that is on the same row as the field of Learning Model ID 302 in which the learning model ID is stored.

[0046] Model User 305 is information about users (other users) who are permitted to use the learning model, identified by the ID stored in the field of Learning Model ID 302. In this embodiment, an example is given where Model User 305 includes User ID 305a and DL permission / failure 305b.

[0047] User ID 305a is the ID of a user who is permitted to use the learning model identified by the ID stored in the field of Learning Model ID 302. In this embodiment, we illustrate a case where the information stored in the field of User ID 305a is determined based on the information entered in the usage permission setting screen described later.

[0048] The DL availability 305b indicates whether or not the learning model identified by the ID stored in the learning model ID 302 field can be downloaded to the learning model user device 120a or 120b of the user (another user) identified by the ID stored in the user ID 305a field. In this embodiment, an example is given in which the content of DL availability 305b is predetermined depending on the intended use of the learning model. For example, if it is expected that the timing of obtaining the inference results by the learning model will be delayed if the learning model is not downloaded and used on the learning model user device 120a or 120b, then DL availability 305b will be "OK" (i.e., it can be downloaded). On the other hand, if it is not expected that such a delay will occur even if the learning model is not downloaded and used on the learning model user device 120a or 120b, then DL availability 305b will be "NG" (i.e., it cannot be downloaded). For example, the processing unit 111 may determine whether or not the learning model can be downloaded depending on the content of content 303. For example, the processing unit 111 may input the contents of content 303 and a prompt including whether or not to output whether or not to download the learning model to the language model 131, thereby obtaining information from the language model 131 indicating whether or not to download the learning model.

[0049] Data provider user 306 is information of another user who provided training data for training the learning model identified by the ID stored in the field of learning model ID 302. In this case, the training data is another person's training data from the perspective of the user identified by the ID stored in the field of user ID 301. In this embodiment, we illustrate a case where data provider user 306 includes user ID 306a and training data ID 306b. In this embodiment, we illustrate a case where the information stored in the fields of data provider user 306 (user ID 306a and training data ID 306b) is determined based on the information entered in the usage permission setting screen described later.

[0050] User ID 306a is the ID of another user who provided training data for training the learning model identified by the ID stored in the field of Learning Model ID 302. The training data ID is the ID of the training data (other users' training data) provided by the user (other user) identified by the ID stored in the field of User ID 306a.

[0051] Profit 307 is the profit obtained from the learning model identified by the ID stored in the field of learning model ID 302. Profit 307 is determined, for example, based on the amount charged for using the learning model. For example, a predetermined amount may be charged each time a learning model is provided. In this case, for example, when the processing unit 111 provides the learning model identified by the ID stored in the field of learning model ID 302, it updates the value of the field in the profit 307 field that is in the same row as the field of learning model ID 302.

[0052] Evaluation 308 is an evaluation of the learning model identified by the ID stored in the field of learning model ID 302. This evaluation may be performed, for example, by a user (another user) to whom the learning model identified by the ID stored in the field of learning model ID 302 has been provided. In this case, for example, when the receiving unit 112 receives the learning model ID and information indicating the evaluation of the learning model identified by that ID from another user (any one of the learning model user devices 120a to 120b) other than the user of the learning model (the local user), the processing unit 111 updates the value of the field in evaluation 308 that is in the same row as the field of learning model ID 302 in which the ID is stored.

[0053] As mentioned above, in this embodiment, the ID of the platform provider of the display control system (the user of the learning model providing device 110), the ID of the learning model of the platform provider, and the main body information of the learning model of the platform provider are also stored in the learning model storage unit 114a. Therefore, in this embodiment, the learning model-related information 310a to 310b includes information about the learning model of the platform provider of the display control system (the user of the learning model providing device 110).

[0054] Furthermore, the contents of each field in the learning model-related information group 300 may be set and updated manually (by input operations on the user interface) by, for example, the user of the learning model providing device 110.

[0055] Returning to the explanation of Figure 2, the learning data storage unit 114c stores, for example, the user IDs of the learning model user devices 120a to 120b, the IDs of the learning data for those users, and the main information of the learning data for those users, each associated with each other. The main information of the learning data is the information that constitutes the learning data, and in this embodiment, it is the learning data to which the correct answer labels have been assigned.

[0056] The learning data-related information storage unit 114d stores learning data-related information for each learning data whose main body information is stored in the learning data storage unit 114c. Learning data-related information is information managed by the learning model provider device 110 as information related to the learning data, separate from the main body information of the learning data. In this embodiment, an example is given in which the learning model-related information includes information necessary to display the usage permission setting screen described later.

[0057] Figure 4 shows an example of the training data-related information group 400. As mentioned above, "***" is a symbol indicating that specific information (numbers, letters, etc.) has been omitted, and "···" is a symbol indicating that similar information may exist above and below it.

[0058] Figure 4 illustrates a case where the training data-related information 410a to 410b includes user ID 401, training data ID 402, content 403, profile 404, data user 405, profit 406, evaluation 407, remaining period 408, and remaining count 409.

[0059] User ID 401 is the ID of the user of the training data (the user). As described above, this embodiment illustrates a case where a user who wishes to create a training model (any one of the training model user devices 120a to 120b) provides training data to be used for training the training model, and transmits the user's ID and name, and the name of the training data, to the training model provider device 110. In this case, for example, the receiving unit 112 receives the user's ID, and the processing unit 111 stores the ID in the user ID 401 field. Figures 3 and 4 illustrate a case where the training model and training data of the same user are associated with each other by the ID stored in user ID 301 of the training model related information 310a to 310b and the ID stored in user ID 401 of the training data related information 410a to 410b.

[0060] The training data ID 402 is the ID of the training data. In this embodiment, we illustrate a case where the processing unit 111 assigns the training data ID to user ID 401 when it stores the user ID of the training data as described above. In this case, the processing unit 111 stores the training data ID 402 in a field of the same row as the field of user ID 401 in which the user ID is stored. In this case, the processing unit 111 may also notify the user of the training data (any one of the training model user devices 120a to 120b) of the user's training data ID via the provision unit 113.

[0061] Content 403 is information that concisely describes the overview of the learning model to be trained using the training data. As mentioned above, this embodiment illustrates a case in which a user who wishes to create a learning model (any one of the learning model user devices 120a to 120b) sends information indicating the content to the learning model provider device 110 when providing the training data to be used for training the learning model. In this case, for example, the processing unit 111 stores this information in a field of content 403 that is in the same row as the field of training data ID 402 in which the ID of the training data is stored. Note that the content of content 303 and 403 is the same for users with the same ID stored in user IDs 301 and 401, so one of content 303 or 403 may be omitted.

[0062] Profile 404 is information that provides a more detailed overview of the learning model to be trained using the training data than the content itself. As mentioned above, this embodiment illustrates a case where a user who wishes to create a learning model (any one of the learning model user devices 120a to 120b) sends information indicating the contents of the profile to the learning model provider device 110 when providing the training data to be used for training the learning model. In this case, for example, the processing unit 111 stores this information in a field of profile 404 that is in the same row as the field of training data ID 402 where the ID of the training data is stored. Note that the contents of profiles 304 and 404 are the same for learning models of users with the same ID stored in user IDs 301 and 401, so one of profiles 304 and 404 may be omitted.

[0063] Data User 405 is information about a user (another user) who is permitted to use the training data, identified by the ID stored in the field of the training data ID 402. In this embodiment, we illustrate the case where Data User 405 includes User ID 405a and Training Model ID 405bb.

[0064] User ID 405a is the ID of a user who is permitted to use the training data, identified by the ID stored in the field of training data ID 402. In this embodiment, we illustrate a case where the information stored in the field of user ID 405a is determined based on the information entered into the usage permission setting screen, which will be described later.

[0065] Learning model ID 405b is the ID of the learning model for the user identified by the ID stored in the field of user ID 405a. Note that the learning model ID is the same for users with the same ID stored in user IDs 301 and 401, so learning model ID 405b is not necessary.

[0066] Profit 406 is the profit obtained from the training data identified by the ID stored in the field of training data ID 402. Profit 406 is determined, for example, based on the amount charged for the use of the training data. In this embodiment, an example is given in which the information stored in the field of profit 406 is determined based on the information entered in the usage permission setting screen described later.

[0067] Evaluation 407 is an evaluation of the training data identified by the ID stored in the field of training data ID 402. This evaluation may be performed, for example, by a user (another user) who is authorized to use the training data identified by the ID stored in the field of training data ID 402. In this case, for example, when the receiving unit 112 receives the ID of the training model and information indicating the evaluation of the training data used for training the training model identified by that ID from another user (any one of the training model user devices 120a to 120b) other than the user of the training data (the local user), the processing unit 111 updates the value of the field in evaluation 407 that is in the same row as the field of training model ID 405b in which that ID is stored.

[0068] The remaining period 408 is the remaining period until the expiration date when the learning data identified by the ID stored in the learning data ID 402 field has a usage period limit. In this embodiment, we illustrate a case where the user who is permitted to use the learning data identified by the ID stored in the learning data ID 402 field is not limited to a specific user (other user), but is a user (other user) who meets the conditions determined based on the information entered in the usage permission setting screen described later, and the remaining period during which such a user can use the learning data is set as the value of remaining period 408. Thus, in this embodiment, we illustrate cases where the value of remaining period 408 is set and cases where it is not set. In this embodiment, we illustrate a case where the initial value of the information stored in the remaining period 408 field is determined based on the information entered in the usage permission setting screen described later.

[0069] The remaining uses of 409 is the number of uses remaining when the learning data identified by the ID stored in the field of learning data ID 402 has a usage limit. In this embodiment, we illustrate a case where the number of times a user can use the learning data identified by the ID stored in the field of learning data ID 402 is not limited to a specific user (other user), but is set to a user (other user) who meets the conditions determined based on the information entered in the usage permission setting screen described later, and the remaining uses of the learning data are set as the value of remaining uses 409. Thus, in this embodiment, we illustrate cases where the value of remaining uses 409 is set and cases where it is not. In this embodiment, we illustrate a case where the initial value of the information stored in the remaining uses 409 field is determined based on the information entered in the usage permission setting screen described later.

[0070] As mentioned above, in this embodiment, the ID of the platform provider of the display control system (the user of the learning model providing device 110), the ID of the learning model of the platform provider, and the main body information of the learning model of the platform provider are also stored in the learning model storage unit 114a. Therefore, in this embodiment, the learning data-related information 410a to 410b is exemplified as including information on the learning data of the platform provider of the display control system (the user of the learning model providing device 110).

[0071] Furthermore, the contents of each field in the learning data-related information group 400 may be set and updated manually (by input operations on the user interface) by the user of the learning model providing device 110, for example.

[0072] [Learning model user devices 120a~120b] The reception unit 122 receives various types of information from external devices. As shown in Figure 1, this embodiment illustrates a case where the reception unit 122 receives information transmitted from an external device via the network 140. This embodiment also illustrates a case where the reception unit 122 functions as a user interface for the learning model user device 120a. In this case, the reception unit 122 receives input operations from the user of the learning model user device 120a.

[0073] The information provision unit 123 provides various types of information to an external device. As shown in Figure 1, this embodiment illustrates a case where the information provision unit 123 transmits information to an external device via the network 140. The processing unit 121 provides overall control of the learning model user device 120a. In this embodiment, the processing unit 121 is shown as performing the creation and management of learning data and the management of the learning model. In this embodiment, the processing unit 121 is shown as having a display control unit 121a. The display control unit 121a performs processing to display the usage permission setting screen, which will be described later, on the display unit 125a.

[0074] The storage unit 124 stores various types of information necessary for processing by the learning model user device 120a. In this embodiment, an example is given in which the storage unit 124 has a learning data storage unit 124a and a learning model storage unit 124b.

[0075] The learning data storage unit 124a stores, for example, the main body information of the learning data used for training the user's learning model on the learning model user device 120a. For example, when the receiving unit 122 receives the main body information of the learning data to which the correct answer label has been assigned, the processing unit 121 may store the main body information of the learning data in the learning data storage unit 124a. Alternatively, the processing unit 121 may perform annotation (assignment of the correct answer label).

[0076] For example, when a user of the learning model user device 120a wishes to create a learning model, they perform an input operation to the learning model user device 120a, inputting their ID and information indicating the content and profile of the learning model. In this case, the receiving unit 122 receives the user's ID and the information indicating the content and profile of the learning model. Also in this case, the processing unit 121 creates information including the user's ID, the information indicating the content and profile of the learning model, and the main body information of the training data used to create the learning model, and the providing unit 123 transmits the information created by the processing unit 121 to the learning model providing device 110.

[0077] The learning model storage unit 124b stores, for example, the main body information of the learning model of a user of the learning model user device 120a. Also, for example, if a user of the learning model user device 120a is permitted to use another user's learning model, the learning model storage unit 124b stores the main body information of that other user's learning model. In this embodiment, the learning model storage unit 124b stores the main body information of a learning model only when both the user's own learning model and another user's learning model can be downloaded from the learning model provider device 110.

[0078] [Usage permission setting screen] Figures 5A and 5B show an example of the permission setting screen 500. In this embodiment, the case where the permission setting screen 500 is a GUI (Graphical User Interface) is illustrated. Furthermore, in this embodiment, the creation of the information necessary for displaying the permission setting screen 500 is performed by the learning model providing device 110 (processing unit 111), the permission setting screen 500 is displayed on the display unit 125a of the learning model user device 120a, and input operations on the permission setting screen 500 are performed by the user of the learning model user device 120a. In the following description, the information necessary for displaying the permission setting screen will be referred to as permission setting screen information as needed.

[0079] Figure 5A shows an example of the initial state of the permission settings screen 500. For example, when the reception unit 122 of the display control unit 121a receives a request from a user of the learning model user device 120a to display the permission settings screen 500, it instructs the provision unit 123 to display the permission settings screen 500. In response to this instruction, the provision unit 123 transmits information indicating the request to display the permission settings screen 500 to the learning model provider device 110. When the reception unit 112 of the learning model provider device 110 receives this information, the processing unit 111 creates the permission settings screen information necessary to display the permission settings screen 500 (initial state) as illustrated in Figure 5A, and instructs the provision unit 113 to transmit the permission settings screen information. In response to this instruction, the provision unit 113 transmits the permission settings screen information to the learning model user device 120a. When the receiving unit 122 receives the usage permission setting screen information, the processing unit 121 (display control unit 121a) of the learning model user device 120a displays the usage permission setting screen 500, as exemplified in Figure 5A, on the display unit 125a based on the usage permission setting screen information. The display unit 125a is implemented using a computer display such as a liquid crystal display or an organic EL (Electro Luminescence) display.

[0080] Figure 5A illustrates a usage permission setting screen 500 for a single training data set for the sake of simplicity. Figure 5A also illustrates a case where the usage permission setting screen 500 displays the following fields: Usage Permission Target 501, Training Model ID 502, Content 503, Profile 504, Evaluation 505, Data Provider 506, Profit 507, Distribution Rate 508, Training Model Users 509, Model Usage Permission 510, Download Permission 511, Permission / Disclosure 512, Disclosure / Confidentiality 513, Period 514, Number of Uses 515, Remarks 516, OK button 517, and Cancel button 518.

[0081] The "Permissioned Users 501" column displays the names of users who are permitted to use the training data. For example, the name of the user with the ID stored in the User ID 301 field will be displayed in the "Permissioned Users 501" column. Figure 5A illustrates a case where the name of the platform provider of the display control system (the user of the training model provider device 110) (see "PF" in Figure 5A) and the names of users who use the display control system (training model user devices 120a to 120b) but are different from the user who performs the input operation on the permission setting screen 500 (see "University X", "Company Y", and "Company Z" in Figure 5A) are displayed in the "Permissioned Users 501" column. Furthermore, as mentioned above, this embodiment illustrates a case where the users permitted to use the training data are not limited to specific users (other users), but can be any user (other user) who meets the conditions determined based on the information entered on the permission setting screen 500. In Figure 5A, "Arbitrary" indicates such a user.

[0082] The Learning Model ID 502 field displays the ID of the learning model of the user (other user) shown in the Permissioned Users 501 field. For example, in the Learning Model ID 302 field, the ID of the user whose name is displayed in the Permissioned Users 501 field is stored in the field of User ID 301, and this ID is stored in the field of Learning Model ID 502. Note that the learning model of the "arbitrary" user displayed in the Permissioned Users 501 field is not specified, so in this embodiment, we will illustrate the case where "-" is displayed in the Learning Model ID 502 field in the same column (row) as the "arbitrary" field ("-" indicates that there is no corresponding information).

[0083] The Content 503 column displays the content of the learning model of the user (other user) shown in the Permissioned Users 501 column. For example, the information stored in the field in the same row as the field in Content 303 that stores the ID of the user whose name is displayed in the Permissioned Users 501 column will be displayed in the Content 303 column.

[0084] Profile 504 displays the learning model profile of a specific user (another user) shown in the "Permissioned Users 501" column. For example, the information stored in the field of Profile 304 that is in the same row as the User ID 301 field, where the ID of the user whose name is displayed in the "Permissioned Users 501" column is stored, will be displayed in Profile 504.

[0085] Furthermore, in this embodiment, we illustrate a case where a dropdown list 504a is displayed in the "Profile 504" field for the learning model of an "Any" user displayed in the "Permissioned Users 501" field, for selecting the profile conditions of the learning model of a user who is permitted to use their own training data. The user of the learning model user device 120a selects the profile conditions of the learning model of a user who is permitted to use their own training data by operating the dropdown list 504a. The profile conditions are, for example, the intended use of the learning model.

[0086] The Evaluation 505 column displays the evaluation of the learning model of a specific user (another user) shown in the Permission to Use 501 column. For example, the information stored in the field on the same row as the User ID 301 field in the Evaluation 308 field, where the ID of the user whose name is displayed in the Permission to Use 501 column is stored, is displayed in the Evaluation 505 column. In Figure 5A, A, B, and C indicate the order of highest evaluation. Also, since the learning model of an "arbitrary" user displayed in the Permission to Use 501 column is not specified, this embodiment illustrates the case where "-" is displayed in the Evaluation 505 column for that "arbitrary" user's learning model ("-" indicates that there is no corresponding information).

[0087] The Data Provider 506 column displays the names of users who provided training data for the learning model of a specific user (another user) shown in the Permissioned Users 501 column, as well as the order in which the training data was provided. For example, in the User ID 306a field of Data Provider User 306, the names of users whose IDs are stored in the field of User ID 301, which is in the same row as the User ID 301 field where the ID of the user whose name is displayed in the Permissioned Users 501 column is stored, are displayed in the Data Provider 506 column from top to bottom in order of provision. Figure 5A shows that the users who provided training data for the learning model of "PF" are "University X" and "Company Y", that "University X" was the first to provide training data, and "Company Y" was the second to provide training data (see "(1) University X (2) Company Y"). Note that Figure 5A also illustrates the case where "-" is displayed to indicate that no users provided training data for the learning model of "University X". Furthermore, in Figure 5A, the learning model of the "arbitrary" user displayed in the column for the user permitted to use 501 is not specified. Therefore, in this embodiment, we illustrate the case where "-" is displayed in the column for the data provider 506 for the learning model of the "arbitrary" user ("-" indicates that there is no corresponding information).

[0088] In this embodiment, the information displayed in the field for data provider 506 is an example of user information that has already been provided. Furthermore, in this embodiment, we illustrate a case where the same name is displayed in multiple fields of the data provider 506 to indicate that the training data of a user with that name is being used to train training models of multiple other users. Figure 5A illustrates a case where "University X" is displayed in three fields of the data provider 506. This indicates that the training data of "University X" is being used to train training models of "PF", "Company Y", and "Company Z". In this embodiment, we illustrate a case where the name of the same user (more specifically "University X") displayed in multiple fields of the data provider 506 is an example of trained training model information.

[0089] The Profit 507 column displays the profit obtained from the learning model of a specific user (another user) shown in the Permission to Use 501 column. For example, the Profit 307 column displays information stored in the field of User ID 301, which is the same row as the User ID field in the Profit 307 column, where the ID of the user whose name is displayed in the Permission to Use 501 column is stored. In this embodiment, the information displayed in the Profit 307 column is an example of the profit obtained from the learning model of a second user.

[0090] Furthermore, in this embodiment, we illustrate a case where a dropdown list 507a for selecting the conditions for the benefits obtained by the learning model of a user who is permitted to use their own training data is displayed in the "Benefits 507" column for the learning model of an "arbitrary" user, which is displayed in the "Permission to Use 501" column. The user of the learning model user device 120a selects the conditions for the benefits obtained by the learning model of a user who is permitted to use their own training data by operating the dropdown list 507a. The benefit conditions are, for example, a lower limit of the amount of benefit.

[0091] The "Distribution Rate 508" column displays the distribution rate of profits earned from the learning model of a specific user (another user) shown in the "Permission to Use 501" column. In this embodiment, the information displayed in the "Profit 507" and "Distribution Rate 508" columns illustrates the compensation that a user can receive when granting permission to use their own training data for training the learning model of a specific user (another user) shown in the "Permission to Use 501" column. Specifically, the compensation is calculated by multiplying the amount displayed in the "Profit 507" column by the distribution rate displayed in the "Distribution Rate 508" column and dividing the result by 100. In this embodiment, the information displayed in the "Distribution Rate 508" column is an example of distribution rate information.

[0092] The distribution rate displayed in Distribution Rate 508 may be a fixed value. However, the distribution rate may be a variable value in order to effectively motivate each user (the user) to allow other users to use their training data. For example, the processing unit 111 may set the distribution rate such that the value of the distribution rate decreases as the number of other users who have provided training data for a certain user's learning model increases. Figure 5A illustrates a case where the distribution rate is 10% when the number of users who have provided training data is 0, and the distribution rate is halved for each increase of 1 in the number of users who have provided training data. In this way, the user of the learning model user device 120a will receive greater compensation the sooner they allow other users to use their training data. Therefore, it is possible to effectively motivate each user (the user) to allow other users to use their training data.

[0093] Furthermore, for example, the processing unit 111 may set the distribution ratio such that the distribution ratio increases as the amount of new data included in the user's training data of the learning model user device 120a increases. New data refers to, for example, data that has different characteristics from the training data of other users. For example, the processing unit 111 may input the user's training data of the learning model user device 120a and a prompt that includes calculating a value indicating the degree of novelty of the training data to the language model 131, thereby obtaining a value indicating the degree of novelty of the training data output from the language model 131. In this case, the processing unit 111 may include in the prompt, for example, setting the minimum value of the value indicating the degree of novelty of the training data to 0. In this case, the processing unit 111 may calculate the distribution ratio so as to be directly proportional to the value indicating the degree of novelty of the training data. Furthermore, the value indicating the degree of novelty of the training data may be calculated by a trained model other than the language model 131, or by other methods. Other methods may include, for example, a method for calculating a value indicating the degree of novelty of the training data based on the difference in feature vectors determined by the features of each training data.

[0094] Furthermore, the processing unit 111 may calculate the distribution ratio based on both the number of users who provided training data for other users' learning models and the degree of novelty of the training data.

[0095] Furthermore, in this embodiment, we illustrate a case where a dropdown list 508a for selecting the distribution rate conditions when allowing an "optional" user to use their own training data is displayed in the distribution rate 508 column for an "optional" user displayed in the permitted users 501 column. In this case, the user of the training model user device 120a selects the distribution rate conditions for themselves when allowing the use of their own training data by operating the dropdown list 508a. The distribution rate conditions are, for example, the lower limit of the distribution rate. In this case, the value obtained by multiplying the profit that satisfies the conditions selected in the profit dropdown list 507a by the distribution rate that satisfies the conditions selected in the distribution rate dropdown list 508a becomes the condition for the compensation that the user receives when using the training data.

[0096] The "Learning Model Users 509" column displays the names of users who are permitted to use the learning model of the specific user (other user) shown in the "Permissioned Users 501" column. For example, Figure 5A shows that the users permitted to use the "PF" learning model are "University X," "Company Y," and "Company Z."

[0097] In Figure 5A, we illustrate the case where "-" is displayed to indicate that no users are permitted to use the learning model of "Company Y". Also, in Figure 5A, the learning model of the "Any" user displayed in the "Permitted Users 501" column is not specified, so in this embodiment, we illustrate the case where "-" is displayed in the "Learning Model Users 509" column, which is in the same column as the "Any" column ("-" indicates that there is no corresponding information). In this embodiment, the information displayed in the "Data Provider 506" column is an example of provided user information.

[0098] The Model Availability 510 column displays information indicating whether the user of the learning model user device 120a can use the learning model of a specific user (another user) displayed in the Permitted Users 501 column. In this embodiment, when permission / denial 512, described later, is set to allow the use of the user's own learning data for training the learning model of a specific user (another user) displayed in the Permitted Users 501 column, an example is shown in the Model Availability 510 column, in the same column as the field where permission is set, indicating that the use of the learning model of that specific user (another user) is permitted. On the other hand, when permission / denial 512, described later, is set to deny permission (denial) for the use of the user's own learning data for training the learning model of a specific user (another user) displayed in the Permitted Users 501 column, an example is shown in the Model Availability 510 column, in the same column as the field where denial is set, indicating that the use of the learning model of that specific user (another user) is not permitted.

[0099] Furthermore, in this embodiment, we illustrate a case where a dropdown list 510a for selecting the conditions of a user who is permitted to use their own training data is displayed in the "Availability of Model Use" column 510 for any user's training model displayed in the "Permission to Use" column 501. The user of the training model user device 120a selects the conditions of the training model that they wish to use by operating the dropdown list 510a. The conditions of the training model that the user wishes to use may be represented, for example, by the owner of the training model (user name).

[0100] The "DL Availability 511" column displays information indicating whether or not the learning models of specific users (other users) shown in the "Permission to Use 501" column can be downloaded. For example, the information stored in the field on the same row as the "User ID 301" field in the "DL Availability 305b" field, where the ID of the user whose name is displayed in the "Permission to Use 501" column is stored, is displayed in the "DL Availability 511" column. Figure 5A illustrates a case where "×" is displayed in the "DL Availability 511" column, indicating that the learning models of all specific users ("PF", "University X", "Company Y", and "Company Z") cannot be downloaded.

[0101] Furthermore, in this embodiment, an example is given in which a dropdown list 511a is displayed in the "DLability 511" column for the learning model of an "arbitrary" user displayed in the "permitted users 501" column, allowing the user to select whether or not they wish for the learning model of a user who is permitted to use their own learning data to be a downloadable learning model. The user of the learning model user device 120a operates the dropdown list 510a to select whether or not they wish for the learning model of a user who is permitted to use their own learning data to be a downloadable learning model.

[0102] The Allow / Deny 512 column allows users to choose whether or not to allow the use of their training data for training the learning models of other users, as indicated in the Allowed Users 501 column. In this embodiment, the drop-down lists 512a to 512e are used to illustrate the case where permission is selected. In this embodiment, the information displayed in the Allow / Deny 512 column is an example of permission and denial information.

[0103] Furthermore, in this embodiment, we illustrate a case where the remarks column 516 of the permission / denial 512 displays text indicating the advantages and disadvantages of allowing one's own training data to be used for training another user's training model.

[0104] Specifically, in the remarks column 516 of approval / denial 512, "The price of your learning model will decrease by XX yen" is an example of information indicating a benefit, and also an example of information indicating that the price of your own learning model will decrease. Note that a decrease in the price of your own learning model means that the amount you pay when training your own learning model using your own training data will decrease.

[0105] The extent to which the price of one's own learning model is reduced (the "XX" in the aforementioned "XX yen") may be a fixed value. However, in order to effectively motivate each user (the user) to allow other users to use their own training data, the extent to which the price of one's own learning model is reduced may be a variable value. For example, the processing unit 111 may input to the language model 131 a prompt that includes calculating the discount amount for the learning model trained using the training data of the user of the learning model user device 120a, and the training data of the user trained using that training data, and obtain the discount amount output from the language model 131. In this case, the processing unit 111 may include in the prompt, for example, that the discount amount is larger the more new data is included in the training data of the user of the learning model user device 120a. Furthermore, the discount amount may be calculated by a trained model other than the language model 131, or by other methods. Such other methods may include, for example, a method of calculating the discount amount based on the difference in feature vectors determined by the features of each training data. In this embodiment, an example of degree information is a specific value (discount amount) that represents the extent to which the price of the learning model is reduced (the "XX" in the aforementioned "XX yen"). Alternatively, instead of a discount amount, a discount rate may be used as the degree information.

[0106] Furthermore, in the remarks column 516 of the permission / denial 512, the statement "Your learning model can also use the learning data of permitted users" is an example of merit information. For example, the more names of users displayed in the learning model users 509 column for a particular learning model, the more users support that learning model. If a user of the learning model user device 120a grants permission to users of that learning model (other users) to use their learning data, they will be able to use the learning data used in that learning model that is supported by many users. Therefore, it is possible to effectively motivate each user (their own user) to grant permission to other users to use their learning data. For example, the learning model "PF" is used by "University X", "Company Y", and "Company Z" (see the learning model users 509 column for the learning model "PF"). Therefore, it is possible to motivate the user of the learning model user device 120a to grant permission to "PF" to use their learning data.

[0107] Furthermore, in the remarks column 516 of the permission / denial 512, the statement "Your learning model cannot use the training data of unauthorized users" is an example of disadvantage information. For example, as mentioned above, the more names of users displayed in the learning model users 509 column for a particular learning model, the more users support that learning model. A user of the learning model user device 120a cannot use the training data used by the learning model supported by many users unless they grant permission to the users (other users) of the learning model supported by many users to use their own training data. Therefore, the same effect as when "Your learning model can also use the training data of authorized users" is displayed is achieved (i.e., it effectively motivates each user to grant permission to other users to use their own training data).

[0108] Furthermore, Figure 5A illustrates a case where the data confirmation button 512f is displayed in the remarks column 516 of the permission / denial 512. When the data confirmation button 512f is pressed, detailed information of the training data is displayed. In this embodiment, an example is shown where the detailed information of the training data is displayed within a single frame. In the following description, this frame will be referred to as the training data viewing screen as needed. Also, in the following description, the information necessary for displaying the training data viewing screen will be referred to as training data viewing screen information as needed.

[0109] Figure 6 shows an example of a learning data viewing screen 600. In this embodiment, the case where the learning data viewing screen 600 is a GUI is illustrated. Furthermore, in this embodiment, the creation of the usage permission setting screen information is performed by the learning model providing device 110 (processing unit 111), the learning data viewing screen 600 is displayed on the display unit 125a of the learning model user device 120a, and input operations are performed on the learning data viewing screen 600 by the user of the learning model user device 120a.

[0110] For example, when the reception unit 122 receives confirmation that the data confirmation button 512f has been pressed, the display control unit 121a instructs the provision unit 123 to request the display of the learning data viewing screen 600. In response to this instruction, the provision unit 123 transmits information indicating the request to display the learning data viewing screen 600 to the learning model providing device 110. When the reception unit 112 receives this information, the processing unit 111 of the learning model providing device 110 creates the learning data viewing screen information necessary for displaying the learning data viewing screen 600 as illustrated in Figure 6, based on the learning data-related information 410a to 410b, and instructs the provision unit 113 to transmit the learning data viewing screen information. In response to this instruction, the provision unit 113 transmits the learning data viewing screen information to the learning model user device 120a. When the receiving unit 122 receives the learning data viewing screen information, the processing unit 121 (display control unit 121a) of the learning model user device 120a displays the learning data viewing screen 600, as illustrated in Figure 6, on the display unit 125a based on the learning data viewing screen information. The learning data viewing screen 600 may, for example, pop up in response to the usage permission setting screen 500.

[0111] Figure 6 illustrates a case where the training data viewing screen 600 displays the user ID 601, training data ID 602, data name 603, content 604, profile 605, training model ID 606, evaluation 607, profit 608, remaining period 609, remaining attempts 610, and an OK button 611.

[0112] The User ID 601 field displays the user ID stored in the User ID 401 field. The "Learning Data ID 602" column displays the learning data ID stored in the field of "Learning Data ID 402" that is in the same row as the field of "User ID 401" where the ID displayed in the "User ID 601" column is stored. The "Data Name 603" field displays the name of the training data identified by the ID shown in the "Training Data ID 602" field.

[0113] The Content 604 column displays information indicating the content stored in the field of User ID 401, which is the same row in Content 403 as the field for User ID 401 where the ID displayed in the User ID 601 column is stored. The Profile 605 column displays information indicating the profile contents stored in the field of User ID 401, which is the same row in the Profile 404 field where the ID displayed in the User ID 601 column is stored.

[0114] The "Training Model ID 606" field displays the ID of the training model stored in the field of "Training Model ID 405b" that is in the same row as the field of "User ID 401" where the ID displayed in the "User ID 601" column is stored. This training model ID is the ID of the training model trained using the training data identified by the ID displayed in the "Training Data ID 602" column.

[0115] The "Evaluation 607" column displays information indicating the evaluation stored in the field corresponding to the User ID 401 field in the "Evaluation 407" column, specifically the field in the same row where the ID displayed in the "User ID 601" column is stored. This evaluation is for the training data identified by the ID displayed in the "Training Data ID 602" column.

[0116] The Profit 608 column displays information indicating the profit stored in the field of Profit 406 that is in the same row as the field for User ID 401, where the ID displayed in the User ID 601 column is stored. This profit is the profit obtained from the training data identified by the ID displayed in the Training Data ID 602 ​​column.

[0117] The remaining period 609 displays information indicating the remaining time until the expiration date, stored in the field of User ID 401, which is the same row as the field for User ID 401 in the field for User ID 601 of the remaining period 408. The remaining usage count of 610 is displayed in the field corresponding to the field for User ID 401, which is the same row in the field for User ID 601 that stores the ID displayed in the field for User ID 401, within the field for Remaining Usage Count 409.

[0118] For example, a user of the learning model user device 120a can refer to the information displayed on the learning data viewing screen 600 to select other users to whom they can allow the use of their own learning data. For example, based on the information displayed in the fields for User ID 601 and Rating 607, they can identify other users who have highly rated learning data. The learning model of that other user is trained using that highly rated learning data. Therefore, it is possible to encourage each user (their own user) to accept the benefits of using that other user's learning model. Thus, it is possible to motivate each user (their own user) to allow other users with highly rated learning data to use their own learning data.

[0119] Similarly, for example, based on the information displayed in the fields for User ID 601 and Used Learning Model ID 606, it is possible to identify other users who possess training data used to train many learning models. Therefore, it is possible to encourage each user (their own user) to believe that there are benefits to using the learning models of those other users. Thus, it is possible to motivate each user (their own user) to ask other users who possess training data used to train many learning models to allow them to use their own training data. Furthermore, for example, based on the information displayed in the fields for User ID 601 and Profit 608, it is possible to identify other users who possess training data from which significant profits have been made. Therefore, it is possible to encourage each user (their own user) to believe that there are benefits to using the learning models of those other users. Thus, it is possible to motivate each user (their own user) to ask other users who possess training data from which significant profits have been made to allow them to use their own training data.

[0120] In this embodiment, when the OK button 611 is pressed, the display of the learning data viewing screen 600 ends and the usage permission setting screen 500 is displayed again.

[0121] Returning to the explanation of Figure 5A, the Disclosure / Confidential 513 column is configured to determine whether or not to disclose to a third party that permission has been granted to another user to use the learning model of that user (another user) displayed in the Permission Target 501 column, when permission has been granted in the Allow / Disallow 512 column (selected). In this embodiment, the case in which the disclosure is selected is illustrated by the dropdown lists 513a to 513e. Furthermore, in this embodiment, the Disclosure / Confidential 513 column can only be configured (selected) when permission has been granted to use the learning model of that user (another user). In Figure 5A, the inability to configure (select) the Disclosure / Confidential 513 column is indicated by displaying the contents of the dropdown lists 513a to 513e in gray.

[0122] In the Period 514 column, the usage period for the training data is set when the user allows the training model of any user (displayed in the Permissioned Users 501 column) to use the training data. In this embodiment, the case in which the usage period is set by the dropdown list 514a is illustrated. As mentioned above, in this embodiment, the case in which such a usage period is not set for a specific user (another user) displayed in the Permissioned Users 501 column is illustrated. Therefore, in this embodiment, in Figure 5A, the case in which "-" is displayed in the Period 514 column for a specific user ("PF", "University X", "Company Y", and "Company Z") is illustrated ("-" indicates that there is no corresponding information).

[0123] In the "Number of Uses 515" column, the number of times the user's training data is used to train the learning model of any user listed in the "Permission to Use 501" column is set. In this embodiment, the case in which the number of uses is set using the dropdown list 515a is illustrated. As mentioned above, in this embodiment, the case in which such a number of uses is not set for a specific user (another user) listed in the "Permission to Use 501" column is illustrated. Therefore, in this embodiment, in Figure 5A, the case in which "-" is displayed in the "Number of Uses 515" column for a specific user ("PF", "University X", "Company Y", and "Company Z") is illustrated ("-" indicates that there is no corresponding information).

[0124] The OK button 517 is pressed by the user of the learning model user device 120a when confirming the input content for the usage permission setting screen 500. When the OK button 517 is pressed, the display control unit 121a terminates the display of the usage permission setting screen 500 and instructs the provision unit 123 to transmit information indicating the input content for the usage permission setting screen 500. In response to this instruction, the provision unit 123 transmits information indicating the input content for the usage permission setting screen 500 to the learning model provider device 110. When the receiving unit 112 of the learning model provider device 110 receives this information, the processing unit 111 updates the learning model related information group 300 and the learning data related information group 400 based on this information. Figure 5B shows an example of the usage permission setting screen 500 in a state where input has been completed (state where input operations have been performed).

[0125] In the example shown in Figure 5B, for example, in the Model Availability column 510, allowing the user's training data from the learning model user device 120a to be used by "PF", "University X", and "Company Y" increases the profit of that user's training data (see Profit 507, Distribution Rate 508, and Allow / Disallow 512). Therefore, in the Profit 406 field, the value of the field in the same row as the User ID 401 field where the user's ID is stored is updated.

[0126] Furthermore, for example, in the Model Usage Permit 510 column, by granting permission for "PF", "University X", and "Company Y" to use the user's training data on the learning model user device 120a, "PF", "University X", and "Company Y" will be able to use the user's training data (see Permit / Disallow 512). Accordingly, three fields are added to the field for data user 405 (User ID 405a and Learning Model ID 405b) in the same row as the field for User ID 401 where the user's ID is stored, and the IDs of "PF", "University X", and "Company Y" and the ID of the learning model are stored in these added fields, respectively.

[0127] Then, when "PF", "University X", and "Company Y" update their learning models using the user's training data, fields are added to the data provider user 306 (user ID 306a and training data ID 306b) in the same row as the field for user ID 301, which stores the user IDs of "PF", "University X", and "Company Y". The user ID and the ID of the user's training data are then stored in these added fields.

[0128] Furthermore, in Figure 5B, the fact that the user's training data from the learning model user device 120a will be used by "University X" and "Company Y" will not be disclosed to any third party (see Disclosure / Confidentiality 513). Therefore, the name of the user will not be displayed in the data provider column 506 for the learning models of "University X" and "Company Y".

[0129] Furthermore, for example, in the Model Usage Permit 510 column, by granting permission for "PF", "University X", and "Company Y" to use the user's training data on the training model user device 120a, the user will be able to use the training models of "PF", "University X", and "Company Y" (see Permit / Disallow 512 and Remarks column 516). Accordingly, fields are added to the Model User 305 (User ID 305a and Download Permit 305b) in the same row as the User ID 301 field where the IDs of "PF", "University X", and "Company Y" are stored, and these added fields store the user's ID and information indicating that downloading is not permitted.

[0130] Additionally, in the field for Remaining Period 408, the period selected by the dropdown list 514a in the Period 514 column is stored (as the initial value of the remaining period) in the field of User ID 401, which stores the ID of a user who meets the conditions displayed in the "Optional" column of Permitted Users 501.

[0131] Additionally, in the field for User ID 401, which stores the ID of a user who meets the conditions displayed in the "Optional" column of Usage Permission Target 501 among the remaining 409 uses, the number of uses selected by the dropdown list 515a in the Usage Count 515 column is stored (as the initial value of the remaining uses).

[0132] Returning to the explanation of Figure 5A, the cancel button 518 is a button pressed by the user of the learning model user device 120a when canceling the input content for the usage permission setting screen 500. When the cancel button 518 is pressed, the display control unit 121a terminates the display of the usage permission setting screen 500 and instructs the provision unit 123 to send information indicating that the input content for the usage permission setting screen 500 is invalid. In response to this instruction, the provision unit 123 sends information indicating that the input content for the usage permission setting screen 500 is invalid to the learning model provider device 110. When the receiving unit 112 of the learning model provider device 110 receives this information, the processing unit 111 does not update the learning model related information group 300 and the learning data related information group 400, as if no input content for the usage permission setting screen 500 had been made.

[0133] Figures 5A and 5B illustrate the permission setting screen 500 for one training data set. In this case, the permission setting screen 500 illustrated in Figures 5A and 5B may be displayed for each training data set. Alternatively, the permission setting screen 500 may be displayed for multiple training data sets. Furthermore, if it is not possible to display all the information at once on the permission setting screen 500, a scroll bar may be displayed on the permission setting screen 500, for example, so that the user of the training model user device 120a can view all the information.

[0134] [Control method for display control system] Next, an example of a control method for the display control system of this embodiment will be described with reference to the flowchart in Figure 7. In this embodiment, an example is given in which the control method for the display control system is realized by using the learning model providing device 110 and the learning model user devices 120a to 120b. For the sake of simplicity, it is assumed that before the processing according to the flowchart in Figure 7 begins, the information that needs to be prepared in advance in order to execute the processing according to the flowchart in Figure 7 is stored in the storage units 114 and 124.

[0135] First, in step S701, the display control unit 121a of the learning model user device 120a determines whether the reception unit 122 has received a request from the user of the learning model user device 120a to display the usage permission setting screen 500. If, as a result of this determination, the reception unit 122 has not received a request to display the usage permission setting screen 500 (the result is NO in step S701), the process according to the flowchart in Figure 7 ends.

[0136] On the other hand, if the reception unit 122 receives a request to display the usage permission setting screen 500 (if the answer is YES in step S701), the process in step S702 is performed. In step S702, the display control unit 121a instructs the provision unit 123 to request the display of the usage permission setting screen 500. In response to this instruction, the provision unit 123 transmits information indicating the request to display the usage permission setting screen 500 to the learning model providing device 110. The reception unit 112 of the learning model providing device 110 receives this information.

[0137] Next, in step S703, the processing unit 111 calculates the allocation rate for each user (learning model) stored in the learning model related information group 300. For example, the allocation rate is calculated based on the number of fields for user ID 305a (the number of other users using the learning model).

[0138] Next, in step S704, the processing unit 111 obtains the discount amount for the user's learning model on the learning model user device 120a. For example, the processing unit 111 instructs the providing unit 113 to input the user's learning data on the learning model user device 120a and a prompt that includes calculating the discount amount for the learning model to be learned using the learning data into the language model 131. Based on this instruction, the providing unit 113 sends the user's learning data and the prompt to the language model server 130. The language model server 130 inputs the user's learning data and the prompt into the language model 131 and sends the discount amount for the user's learning model output from the language model 131 to the learning model provider device 110. The receiving unit 112 receives the discount amount for the user's learning model and outputs it to the processing unit 111. In this way, the processing unit 111 obtains the discount amount for the user's learning model.

[0139] Next, in step S705, the processing unit 111 creates usage permission setting screen information based on the learning model-related information 310a to 310b and the learning data-related information 410a to 410b, the distribution rate calculated in step S703, and the discount amount obtained in step S704, and instructs the provisioning unit 113 to transmit the usage permission setting screen information. In response to this instruction, the provisioning unit 113 transmits the usage permission setting screen information to the learning model user device 120a. The receiving unit 122 of the learning model user device 120a receives the information.

[0140] Next, in step S706, the display control unit 121a displays the permission setting screen 500 on the display unit 125a based on the permission setting screen information received by the reception unit 122 in step S706 (see Figure 5A). If the process in step S706 is performed immediately after the completion of step S705, the permission setting screen 500 in its initial state will be displayed in step S706.

[0141] Next, in step S707, the display control unit 121a determines whether or not the OK button 517 has been pressed. If the result of this determination is that the OK button 517 has been pressed (YES in step S707), the process in step S708 is performed. In step S708, the display control unit 121a instructs the provisioning unit 123 to transmit information indicating the input content to the usage permission setting screen 500 at the time the OK button 517 was pressed. In response to this instruction, the provisioning unit 123 transmits information indicating the input content to the usage permission setting screen 500 to the learning model providing device 110. When the receiving unit 112 of the learning model providing device 110 receives this information, the processing unit 111 updates the learning model related information group 300 and the learning data related information group 400 based on this information.

[0142] Next, in step S709, the display control unit 121a terminates the display of the usage permission setting screen 500. Once the process in step S709 is completed, the process shown in the flowchart in Figure 7 is completed.

[0143] If the result of the determination in step S707 is that the OK button 517 has not been pressed (the result in step S707 is NO), the process in step S710 is performed. In step S710, the display control unit 121a determines whether or not the data confirmation button 512f has been pressed. If the result of this determination is that the data confirmation button 512f has been pressed (the result in step S710 is YES), the process in step S711 is performed. In step S711, the display control unit 121a instructs the provision unit 123 to request the display of the learning data viewing screen 600. In response to this instruction, the provision unit 123 transmits information indicating the request to display the learning data viewing screen 600 to the learning model providing device 110. The receiving unit 112 of the learning model providing device 110 receives this information.

[0144] Next, in step S712, the processing unit 111 of the learning model user device 120a creates learning data viewing screen information based on the learning data-related information 410a to 410b, and instructs the providing unit 113 to transmit the learning data viewing screen information. The providing unit 113 transmits the learning data viewing screen information to the learning model user device 120a in response to this instruction. The receiving unit 122 of the learning model user device 120a receives the learning data viewing screen information.

[0145] Next, in step S713, the display control unit 121a displays the learning data viewing screen 600 on the display unit 125a based on the learning data viewing screen information received by the reception unit 122 in step S712.

[0146] Next, in step S714, the display control unit 121a waits until the OK button 611 is pressed. When the OK button 611 is pressed (if the answer in step S714 is YES), the process in step S715 is performed. In step S715, the display control unit 121a terminates the display of the learning data viewing screen 600. Then, the process in step S706 described above is performed. If the process in step S706 is performed immediately after the completion of step S715, in step S706, the usage permission setting screen 500 that was displayed when the data confirmation button 512f was pressed in step S710 is displayed again.

[0147] If the result of the determination in step S710 is that the data confirmation button 512f has not been pressed (the result in step S710 is NO), the process in step S716 is performed. In step S716, the display control unit 121a determines whether or not the cancel button 518 has been pressed. If the result of this determination is that the cancel button 518 has not been pressed (the result in step S716 is NO), the process in step S707 is performed. On the other hand, if the cancel button 518 has been pressed (the result in step S716 is YES), the process in step S709 is performed.

[0148] [summary] As described above, in this embodiment, the display control system (display control unit 121a) displays information on the display unit 125a indicating permission to allow another user (second user) to use the user's (first user's) training data for training the learning model (see permission / denial 512). At this time, the display control system displays the information indicating permission and the benefit information indicating the benefits of permission on the display unit 125a in a corresponding manner.

[0149] For example, the display control system displays information indicating that compensation can be obtained for training data (see profit 507 and distribution rate 508), information indicating that other users' training models can be used (see model availability 510), information indicating that the price of one's own training model will be reduced (see remarks column 516), and information indicating that other users' training data can be used to train one's own training model (see remarks column 516) as examples of displaying benefit information.

[0150] Therefore, it becomes easier to motivate users (the user themselves) to allow other users to use their training data for machine learning. Furthermore, by displaying information about the benefits, it becomes possible to ensure that users (the user themselves) who decide whether or not to allow other users to use their training data understand the advantages of doing so.

[0151] Furthermore, by displaying information indicating that users can use other users' training models to train their own training models, the benefit of being able to use other users' training models can be made clear to the user (the user themselves) when deciding whether or not to allow the use of training data.

[0152] Furthermore, by displaying information indicating that other users' training data can be used to train one's own learning model, the benefit of being able to train one's own learning model with other users' training data can be made clear to the user (the user themselves) who decides whether or not to allow the use of the training data.

[0153] Furthermore, by displaying information indicating that compensation can be obtained for the training data, the user (the user) deciding whether or not to allow the use of the training data can be made aware that obtaining compensation for their own training data is a benefit. In this case, by using a distribution of profits obtained from other users' training models as compensation for their own training data, the user (the user) deciding whether or not to allow the use of the training data can be made aware that they will receive compensation commensurate with the profits obtained from other users' training models. In addition, by displaying the profit distribution rate obtained from other users' training models (see distribution rate 508), the user (the user) deciding whether or not to allow the use of the training data can be made to quantitatively understand how much of the profits obtained from other users' training models will be distributed to them. Furthermore, by setting the profit sharing rate for other users' learning models to decrease as the number of users who have already provided training data to other users' learning models increases, users (their own users) who decide whether or not to allow the use of their training data can be made aware that by granting permission for other users (or their learning models) to use their own training data sooner, they will receive a larger share of the profits earned by other users' learning models. Additionally, by setting the profit sharing rate for other users' learning models to increase the more new data is included in their own training data, users (their own users) who decide whether or not to allow the use of their own training data can be made aware that the greater the value of their own training data, the larger the profits earned by other users' learning models will be received.

[0154] Furthermore, by displaying information indicating that the price of their own learning model will be reduced, the user (the user themselves) who decides whether or not to allow the use of the training data can be made aware that the reduction in the price of their own learning model is a benefit. In this case, by displaying information indicating the extent to which the price of their own learning model will be reduced, the user (the user themselves) can be made to understand quantitatively how much the price of their own learning model will be reduced.

[0155] Furthermore, in this embodiment, the display control system displays the information in the "Data Provider 506" column on the display unit 125a as an example of provided user information indicating users who have already provided training data for training other users' learning models. Therefore, the user (the own user) who decides whether or not to permit the use of training data can be made aware of the users who have provided training data to other users' learning models. Thus, for example, the value of other users' learning models can be made known to the user (the own user) who decides whether or not to permit the use of training data. For example, the more users who have provided training data to other users' learning models, the higher the value of that learning model can be made known to the user (the own user) who decides whether or not to permit the use of training data.

[0156] Furthermore, in this embodiment, the display control system displays the name of the same user in multiple fields of the data provider 506 on the display unit 125a as an example of trained training model information showing multiple training models trained using other users' training data (see "University X"). Therefore, the user (the own user) who decides whether or not to allow the use of training data can be made aware of multiple training models trained using other users' training data. This allows the user (the own user) who decides whether or not to allow the use of training data to be made aware of, for example, the value of other users' training data. For example, in the examples shown in Figures 5A and 5B, the user (the own user) who decides whether or not to allow the use of training data can be made aware that the training model of "University X" is of high value.

[0157] Furthermore, in this embodiment, the display control system (display control unit 121a) displays information on the display unit 125a indicating that it does not permit the use of the user's (first user's) training data for training another user's (second user's) training model (see permission / denial 512). In this case, the display control system displays the information indicating that permission is denied and the disadvantage information indicating the disadvantages of not granting permission on the display unit 125a in correspondence with each other. For example, the display control system displays information indicating that it cannot use another user's training data for training its own training model (see remarks column 516) as an example of displaying disadvantage information.

[0158] Therefore, similar to displaying benefits, it becomes easier to motivate the user (the user themselves) to allow other users to use the training data used for machine learning. Furthermore, by displaying disadvantages, the user (the user themselves) can be sure to understand the disadvantages of not allowing other users to use their training data in their learning models.

[0159] (Other embodiments) The embodiments of this disclosure described above can be implemented by a computer executing a program. Furthermore, a computer-readable recording medium on which the program is stored, and a computer program product such as the program itself, can also be applied as embodiments of this disclosure. Examples of recording media include flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, magnetic tapes, non-volatile memory cards, ROMs, etc. Moreover, the embodiments of this disclosure may be implemented by a PLC (Programmable Logic Controller) or by dedicated hardware such as an ASIC (Application Specific Integrated Circuit). Furthermore, the embodiments of this disclosure described above are merely examples of concrete implementations of this disclosure, and the technical scope of this disclosure should not be interpreted as being limited by them. In other words, this disclosure can be implemented in various ways without departing from its technical concept or its main features.

[0160] Furthermore, the disclosure of the above embodiments is as follows, for example. [Disclosure 1] A display control system comprising a reception unit for receiving first learning data from a first user, Display control means that causes permission information to be displayed on the display unit indicating permission to use the first training data for training a second user's learning model, which is different from the first user. Equipped with, The display control means causes the permission information and the benefit information indicating the benefits of granting the permission to be displayed on the display unit in a corresponding manner. A display control system characterized by the following features. [Disclosure 2] The aforementioned benefits information includes information indicating that the second user's learning model becomes available to the first user. The display control system according to disclosure 1, characterized in that [Disclosure 3] The display control means causes the display unit to display provided user information indicating users who have already provided training data for training the second user's learning model. The display control system according to disclosure 2, characterized in that [Disclosure 4] The reception unit receives the second learning data from the second user. The aforementioned merit information includes information indicating that the second training data can be used for training the first user's learning model. A display control system according to any one of disclosures 1 to 3, characterized in that [Disclosure 5] The display control means causes the display unit to display trained model information, which shows a plurality of trained models trained using the second training data. The display control system according to disclosure 4, characterized by the features described herein. [Disclosure 6] The aforementioned merit information includes information indicating that compensation can be obtained for the first training data, A display control system according to any one of disclosures 1 to 5, characterized in that [Disclosure 7] The aforementioned consideration is a distribution of the profits obtained from the second user's learning model. The display control system according to disclosure 6, characterized in that [Disclosure 8] The display control means causes the display unit to display distribution rate information indicating the profit distribution rate. The display control system according to disclosure 7, characterized by the features described herein. [Disclosure 9] The aforementioned distribution ratio decreases as the number of users who have already provided training data to the second user's learning model increases. The display control system according to disclosure 8, characterized in that [Disclosure 10] The aforementioned distribution ratio increases as the amount of new data included in the first training data increases. The display control system according to disclosure 9, characterized in that [Disclosure 11] The aforementioned benefit information includes information indicating that the price of the first user's learning model will be lower. A display control system according to any one of disclosures 1 to 10, characterized in that it is a display control system. [Disclosure 12] The display control means causes the display unit to display information indicating the extent to which the price of the first user's learning model will be reduced, as information indicating that the price of the first user's learning model will be reduced. The aforementioned degree information is information output by the language model. The display control system according to disclosure 11, characterized in that [Disclosure 13] The aforementioned first training data includes training data to be used to train the learning model. A display control system according to any one of disclosures 1 to 12, characterized by the features described herein. [Disclosure 14] A display control system comprising a reception unit for receiving first learning data from a first user, Display control means that causes a display unit to display disallowance information indicating that the use of the first training data for training a second user's learning model, which is different from the first user, is not permitted. Equipped with, The display control means causes the display unit to display the disallowance information and the disadvantage information indicating the disadvantages of not allowing the disallowance in a corresponding manner. A display control system characterized by the following features. [Disclosure 15] The reception unit receives the second learning data from the second user. The aforementioned disadvantage is that the second user's second training data is not used for training the first user's learning model. The display control system according to disclosure 14, characterized by the features described herein. [Disclosure 16] A control method for a display control system comprising a reception unit that receives first learning data from a first user, A display control step in which permission information indicating permission to use the first training data for training a second user's learning model, which is different from the first user, is displayed on the display unit. Equipped with, In the display control step, the permission information and the benefit information indicating the benefits of granting the permission are associated with each other and displayed on the display unit. A control method for a display control system characterized by the following: [Disclosure 17] A control method for a display control system comprising a reception unit that receives first learning data from a first user, A display control step in which the display unit displays disallowance information indicating that it is not permitted to use the first training data for training a second user's learning model, which is different from the first user. Equipped with, In the display control step, the denial information and the disadvantage information indicating the disadvantages of not granting permission are displayed on the display unit in a corresponding manner. A control method for a display control system characterized by the following: [Disclosure 18] A program for causing a computer to function as a display control means for the display control system described in any one of disclosures 1 to 15. [Explanation of symbols]

[0161] 110 Learning Model Provider Device 111 Processing Unit 112 Reception Department 113 Provision Department 114 Storage section 114a Learning Model Memory Unit 114b Learning Model Related Information Storage Unit 114c Learning data storage unit 114d Learning data-related information storage unit 120a~120b Learning model user device 121 Processing Unit 121a Display Control Unit 122 Reception Department 123 Provision Department 124 Storage section 124a Learning data storage unit 124b Learning Model Memory Unit 125a~125b Display section 130 Language Model Servers 131 Language Models 140 Networks 300 Learning Model-Related Information Groups 301 User ID 302 Learning Model ID 303 Contents 304 Profile 305 Model Users 305a User ID 305b DL availability 306 Data Providers 306a User ID 306b Training data ID 307 Profit 308 ratings 310a~310b Learning Model Related Information 400 Data related information for training 401 User ID 402 Training Data ID 403 Content 404 Profile 405 Data User 405a Learning Model ID 405b DL availability 406 Profit 407 ratings 408 Remaining period 409 remaining attempts 410a~410b Information related to training data 500 Usage permission setting screen 501 Permitted for Use 502 Learning Model ID 503 Contents 504 Profile 504a Dropdown list 505 ratings 506 Data Providers 507 Profit 507a Dropdown list 508 distribution rate 508a Dropdown list 509 Users of the Learning Model 510 Model Availability 510a Dropdown List 511 DL availability 512 Allow / Disallow 512a~512e Dropdown List 512f Data Confirmation Button 513 Disclosure / Confidentiality 513a~513e Dropdown List 514 period 514a Dropdown list 515 Number of uses 515a Dropdown list 516 Remarks column 517 OK button 518 Cancel button 600 Learning Data Viewing Screen 601 User ID 602 Training Data ID 603 Data Name 604 Contents 605 Profile 606 Used training model ID 607 ratings 608 Profit 609 Remaining period 610 remaining attempts 611 OK button

Claims

1. A display control system comprising a reception unit for receiving first learning data from a first user, Display control means for displaying permission information on a display unit indicating permission to use the first training data for training a learning model of a second user different from the first user, Equipped with, The display control means causes the permission information and the benefit information indicating the benefits of granting the permission to be displayed on the display unit in a corresponding manner. A display control system characterized by the following features.

2. The aforementioned benefits information includes information indicating that the second user's learning model becomes available to the first user. The display control system according to claim 1, characterized in that it is as described above.

3. The display control means causes the display unit to display provided user information indicating users who have already provided training data for training the second user's learning model. The display control system according to claim 2, characterized in that it is as follows.

4. The reception unit receives the second learning data from the second user. The aforementioned merit information includes information indicating that the second training data can be used for training the first user's learning model. The display control system according to claim 1, characterized in that it is as described above.

5. The display control means causes the display unit to display trained model information, which shows a plurality of trained models trained using the second training data. The display control system according to claim 4, characterized in that it is as described above.

6. The aforementioned merit information includes information indicating that compensation can be obtained for the first training data. The display control system according to claim 1, characterized in that it is as described above.

7. The aforementioned consideration is a distribution of the profits obtained from the second user's learning model. The display control system according to claim 6, characterized in that way.

8. The display control means causes the display unit to display distribution rate information indicating the profit distribution rate. The display control system according to claim 7.

9. The aforementioned distribution ratio decreases as the number of users who have already provided training data to the second user's learning model increases. The display control system according to claim 8, characterized in that it is as follows.

10. The aforementioned distribution ratio increases as the amount of new data included in the first training data increases. The display control system according to claim 9.

11. The aforementioned benefit information includes information indicating that the price of the first user's learning model will be lower. A display control system according to any one of claims 1 to 10, characterized by the above.

12. The display control means causes the display unit to display information indicating the extent to which the price of the first user's learning model will be reduced, as information indicating that the price of the first user's learning model will be reduced. The aforementioned degree information is information output by the language model. The display control system according to claim 11, characterized in that it is as described above.

13. The first training data includes training data to be used to train the learning model. A display control system according to any one of claims 1 to 10, characterized by the above.

14. A display control system comprising a reception unit for receiving first learning data from a first user, Display control means that causes a display unit to display disallowance information indicating that the use of the first training data for training a second user's learning model, which is different from the first user, is not permitted. Equipped with, The display control means causes the display unit to display the disallowance information and the disadvantage information indicating the disadvantages of not allowing the disallowance in a corresponding manner. A display control system characterized by the following features.

15. The reception unit receives the second learning data from the second user. The aforementioned disadvantage is that the second user's second training data is not used for training the first user's learning model. The display control system according to claim 14, characterized in that it is as described above.

16. A control method for a display control system comprising a reception unit that receives first learning data from a first user, Display control step of displaying permission information on a display unit indicating permission to use the first training data for training a second user's learning model, which is different from the first user. Equipped with, In the display control step, the permission information and the benefit information indicating the benefits of granting the permission are associated with each other and displayed on the display unit. A control method for a display control system characterized by the following:

17. A control method for a display control system comprising a reception unit that receives first learning data from a first user, A display control step in which the display unit displays disallowance information indicating that it is not permitted to use the first training data for training a second user's learning model, which is different from the first user. Equipped with, In the display control step, the denial information and the disadvantage information indicating the disadvantages of not granting permission are displayed on the display unit in a corresponding manner. A control method for a display control system characterized by the following:

18. A program for causing a computer to function as a display control means for the display control system according to claim 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or 14.

Images