Information processing methods

Abstract

To provide incentives to a pre-trained model provider.SOLUTION: An information processing method includes: an acceptance step of accepting a request to use a pre-trained model registered on a distributed ledger; an issuance step of issuing a token to a wallet of a provider of the pre-trained model when the request is accepted; and a granting step of granting a non-transferable token to the wallet on the basis of a cumulative amount of tokens issued to the wallet.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to the technical field of information processing methods using distributed ledger technology.

Background Art

[0002] As a method of this kind, for example, a method of managing product development data using distributed ledger technology has been proposed (see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] For example, by registering data in a distributed ledger, the registered data can be shared by multiple persons while preventing tampering of the registered data. For example, when a learned model as an output of machine learning is registered in a distributed ledger, a person different from the person who provided the learned model can also use the learned model. On the other hand, the person who provided the learned model is often not given an incentive.

[0005] The present invention has been made in view of the above circumstances, and an object thereof is to provide an information processing method capable of giving an incentive to a provider of a learned model.

Means for Solving the Problems

[0006] An information processing method according to one aspect of the present invention includes: a receiving step of receiving a request to use a trained model registered in a distributed ledger; an issuing step of issuing tokens to the wallet of the provider of the trained model when the request to use is received; and a granting step of granting non-transferable tokens to the wallet based on the cumulative amount of tokens issued to the wallet. [Brief explanation of the drawing]

[0007] [Figure 1] This is a conceptual diagram illustrating the concept of an information processing system according to the embodiment. [Figure 2] This figure shows an example of the configuration of an information processing system according to the embodiment. [Figure 3] A flowchart illustrating an example of the operation of the information processing system according to the embodiment. [Figure 4] A flowchart illustrating another example of the operation of the information processing system according to the embodiment. [Modes for carrying out the invention]

[0008] Embodiments of the information processing method will be described with reference to Figures 1 to 4. An information processing system to which the information processing method according to the embodiment is applied will be described.

[0009] (Concept of Information Processing System 1) First, the concept of the information processing system 1 will be explained with reference to Figure 1. In Figure 1, the information processing system 1 comprises a distributed network 10 and a data management system 20. The distributed network 10 is a network for realizing a distributed ledger. In this embodiment, blockchain is given as an example of a distributed ledger.

[0010] The data management system 20 includes a database (DB) 220. For example, the database 220 may contain trained models generated by machine learning. For example, the trained models may be trained models applicable to an autonomous driving system for a vehicle. For example, the trained models may be trained models applicable to a navigation system. However, the trained models are not limited to trained models applicable to at least one of the autonomous driving system and the navigation system.

[0011] Furthermore, in addition to or instead of the trained model, the database 220 may also contain, for example, at least one of the training data used to train the trained model and the source code used to train the trained model. Furthermore, the data management system 20 may include, in addition to or instead of the database 220 in which the trained model is registered, at least one of the database in which the training data is registered and the database in which the source code is registered.

[0012] The trained models to be registered in the database 220 may be provided by provider P. Provider P may access the information processing system 1 via terminal device 30. For example, provider P may generate a transaction via terminal device 30 requesting the registration of a trained model. For example, the generated transaction may include identification information for identifying provider P and model information indicating a trained model. An example of identification information for identifying provider P is the user account related to provider P. An example of model information is the file name of a trained model.

[0013] As a result of the above-generated transaction being processed in the distributed network 10 (for example, as a result of the transaction being registered in the distributed ledger), provider P may be provided with information for registering a trained model in the database 220. Based on this information, provider P may register a trained model in the database 220 via the terminal device 30. In this case, for example, a trained model may be registered in the database 220 linked to identification information relating to provider P. As an example of information for registering a trained data in the database 220, link information to the database 220 can be cited.

[0014] When a pre-trained model is registered in the database 220, a person other than the provider P (for example, user U) will be able to use the pre-trained model. User U may access the information processing system 1 via the terminal device 40. For example, if user U wishes to use a pre-trained model, user U may generate a transaction 41 via the terminal device 40 requesting the use of the pre-trained model. For example, transaction 41 may include identification information for identifying user U and model information indicating the pre-trained model. An example of identification information for identifying user U is the user account related to user U.

[0015] As a result of transaction 41 being processed in the distributed network 10 (for example, as a result of transaction 41 being registered in the distributed ledger), user U may be provided with information for obtaining a trained model from database 220. Based on this information, user U may obtain a trained model (see symbol "21" in Figure 1) from database 220 via terminal device 40. An example of information for obtaining a trained model from database 220 is information linking to database 220.

[0016] The decentralized network 10 that processed transaction 41 may issue tokens 11 to provider P, who provided a trained model. The tokens 11 may be added to provider P's wallet 31. The wallet 31 may be managed by a distributed ledger implemented by the decentralized network 10, or by a distributed ledger implemented by a decentralized network different from the decentralized network 10.

[0017] The decentralized network 10 may further grant non-transferable tokens 12 to the provider P's wallet 31 based on the cumulative amount of tokens 11 issued to the provider P. For example, if the cumulative amount of tokens 11 issued to the provider P is "1" or more, the decentralized network 10 may grant a token representing the first rank as a non-transferable token 12 to the wallet 31. If the cumulative amount of tokens 11 issued to the provider P is "10" or more, the decentralized network 10 may grant a token representing the second rank, which is a higher rank than the first rank, as a non-transferable token 12 to the wallet 31. If the cumulative amount of tokens 11 issued to the provider P is "100" or more, the decentralized network 10 may grant a token representing the third rank, which is a higher rank than the second rank, as a non-transferable token 12 to the wallet 31. An example of a non-transferable token 12 is the Soulbound Token (SBT).

[0018] (An example of Information Processing System 1) Next, a specific example of the configuration of the information processing system 1 will be described with reference to Figure 2. In Figure 2, the information processing system 1 comprises a distributed network 10 and a data management system 20. The data management system 20 comprises a management server 210 and a database 220. Terminal devices 30 and 40 may each be connected to the management server 210 via a wide-area network such as the Internet.

[0019] The management server 210 may provide an application for viewing a distributed ledger realized by the distributed network 10 and an application for accessing the database 220 to the provider P and the user U. For this reason, the management server 210 may be referred to as an application server.

[0020] For example, when the provider P registers a certain learned model in the database 220, the provider P may transmit registration request information for requesting the registration of the certain learned model to the management server 210 using the application provided by the management server 210. The management server 210 that has received the registration request information may generate a transaction for requesting the registration of the certain learned model. After the transaction is processed in the distributed network 10, the management server 210 may transmit display information for displaying information for registering the certain learned model in the database 220 on the screen related to the application to the terminal device 30. The terminal device 30 that has received the display information may display the information for registering the certain learned model in the database 220. Thereafter, the provider P may register the certain learned model in the database 220 using the application provided by the management server 210.

[0021] For example, when user U uses a single learned model, user U may transmit usage request information for requesting the use of the single learned model to management server 210 using an application provided by management server 210. Management server 210 that has received the usage request information may generate a transaction for requesting the use of the single learned model (for example, corresponding to transaction 41 shown in FIG. 1). After the transaction is processed in distributed network 10, management server 210 may transmit display information for displaying information for acquiring the single learned model from database 220 to terminal device 40 on a screen related to the application. Terminal device 40 that has received the display information may display information for acquiring the single learned model from database 220. Thereafter, user U may acquire the single learned model from database 220 using an application provided by management server 210.

[0022] Distributed network 10 that has processed a transaction for requesting the use of a single learned model may issue token 11 to provider P. In this case, management server 210 may grant token 11 to wallet 31 of provider P. Distributed network 10 may further issue non-transferable token 12 to provider P based on the cumulative amount of token 11 issued to provider P. In this case, management server 210 may grant non-transferable token 12 to wallet 31 of provider P.

[0023] Next, the operation of information processing system 1 will be further described with reference to the flowcharts of FIGS. 3 and 4. In FIG. 3, information processing system 1 (for example, management server 210) receives a request from user U to use a single learned model (step S101). Information processing system 1 (for example, management server 210) provides user U with a single learned model (that is, data) based on the received usage request (step S102). Information processing system 1 (for example, distributed network 10) issues token 11 to provider P who has provided the single learned model in parallel with the processing of step S102 (step S103).

[0024] In Figure 4, the information processing system 1 (e.g., the distributed network 10) determines whether the cumulative amount of tokens 11 issued to provider P is equal to or greater than a threshold (step S201). If it is determined that the cumulative amount of tokens 11 is equal to or greater than a threshold (step S201: Yes), the information processing system 1 (e.g., the distributed network 10) issues non-transferable tokens 12 to provider P (step S202). On the other hand, if it is determined that the cumulative amount of tokens 11 is not equal to or greater than a threshold (step S201: No), the operation shown in Figure 4 is terminated. Note that the operation shown in Figure 4 may be performed when new tokens 11 are issued to provider P.

[0025] (Technical effects) In information processing system 1, when a pre-trained model provided by provider P is used by user U, a token 11 is issued to provider P. Therefore, the more people who use a pre-trained model, the greater the cumulative amount of tokens 11 issued to provider P. For example, the greater the cumulative amount of tokens 11 issued to provider P, the greater the contribution of the pre-trained model provided by provider P to others (e.g., user U). Thus, the cumulative amount of tokens 11 issued to provider P can be said to represent at least one of the evaluation of provider P and the achievements of provider P.

[0026] In the information processing system 1, non-transferable tokens 12 are further issued to provider P based on the cumulative amount of tokens 11 issued to provider P. If the cumulative amount of tokens 11 is relatively large, the information processing system 1 may issue non-transferable tokens 12 of a higher rank to provider P compared to when the cumulative amount of tokens 11 is relatively small. Therefore, it can be said that the non-transferable tokens 12 issued to provider P represent at least one of the evaluation of provider P and the achievements of provider P. Because the non-transferable tokens 12 are non-transferable, the reliability of at least one of the evaluation and achievements of provider P represented by the non-transferable tokens 12 can be guaranteed.

[0027] Furthermore, the non-transferable token 12 may be made public. In this case, user U can refer to provider P's non-transferable token 12 to decide whether or not to use a pre-trained model provided by provider P. Furthermore, token 11 may be a transferable token. For example, token 11 may be a token that can be exchanged for at least one of goods and services.

[0028] Thus, in the information processing system 1, when a trained model is used by user U, token 11 is issued to provider P who provided the trained model, and non-transferable tokens 12 are issued based on the cumulative amount of tokens 11. Therefore, the issuance of tokens 11 and non-transferable tokens 12 can serve as an incentive for provider P to be more willing to provide trained models. Accordingly, the information processing system 1, that is, the information processing method applied to the information processing system 1, can provide an incentive to providers of trained models.

[0029] (modified version) Information processing system 1 (for example, a distributed network 10) may issue one token 11 each time a trained model is used. In other words, information processing system 1 may issue a predetermined amount of tokens 11 for each use of a trained model. Information processing system 1 (for example, a distributed network 10) may issue additional tokens 11 to provider P, who provided the trained model, based on a request from user U. For example, user U, who has used a trained model, may evaluate the trained model. Then, user U may request information processing system 1 to issue additional tokens 11 to provider P according to the evaluation of the trained model. In this case, user U may be required to pay provider P a monetary amount corresponding to the amount of additional tokens 11 issued. In other words, there may be a charge for the additional tokens 11 issued to provider P. Information processing system 1 may constitute at least part of an open data marketplace.

[0030] Various aspects of the invention derived from the embodiments and modifications described above are described below.

[0031] An information processing method according to one aspect of the invention includes: a receiving step of receiving a request to use a trained model registered in a distributed ledger; an issuing step of issuing tokens to the wallet of the provider of the trained model when the request to use is received; and a granting step of granting non-transferable tokens to the wallet based on the cumulative amount of tokens issued to the wallet.

[0032] In this information processing method, the non-transferable token may represent at least one of the provider's evaluation and / or merit.

[0033] An information processing system according to one aspect of the invention comprises a distributed network for realizing a distributed ledger and a management device for managing a trained model registered in the distributed ledger. The distributed network receives requests for the use of the trained model, and upon receiving such a request, issues tokens to the wallet of the provider of the trained model, and grants non-transferable tokens to the wallet based on the cumulative amount of tokens issued to the wallet. The "data management system 20" in the above-described embodiment corresponds to an example of the "management device".

[0034] In the information processing system, the non-transferable token may represent at least one of the provider's evaluation and / or achievements.

[0035] The present invention is not limited to the embodiments described above, and can be modified as appropriate without contradicting the gist or idea of ​​the invention as can be read from the claims and specification as a whole. Information processing methods involving such modifications are also included within the technical scope of the present invention. [Explanation of symbols]

[0036] 1…Information processing system, 10…Distributed network, 20…Data management system, 30, 40…Terminal device, 210…Management server, 220…Database

Claims

1. A receiving process for receiving requests to use a pre-trained model registered in a distributed ledger, When the aforementioned request for use is accepted, the issuance process involves issuing a token to the wallet of the provider of the first trained model, A granting step of granting non-transferable tokens to the wallet based on the cumulative amount of tokens issued to the wallet, An information processing method characterized by including

2. The information processing method according to claim 1, characterized in that the non-transferable token represents at least one of the provider's evaluation and achievements.

Images