Data management methods
A data management method using distributed ledger databases and third-party verifiable certifications addresses the vulnerability of machine learning models to tampered datasets, ensuring reliable and high-quality training data.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- TOYOTA JIDOSHA KK
- Filing Date
- 2024-12-06
- Publication Date
- 2026-06-18
AI Technical Summary
The accuracy of machine learning models is dependent on the learning dataset, and they are vulnerable to data poisoning attacks through tampered datasets.
A data management method involving a management step that utilizes a third-party verifiable proof system, where training data is stored in a storage device and managed using a distributed ledger database, ensuring the data meets predetermined conditions and issues verifiable certifications.
This method ensures the reliability of training data by preventing tampering and providing high-quality, legitimate data through third-party verifiable certifications.
Smart Images
Figure 2026099169000001_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data management methods.
Background Art
[0002] For example, as learning of an AI (Artificial Intelligence) model, learning using a learning dataset that is a group of teacher data including learning data and correct labels 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, the accuracy of machine learning often depends on the learning dataset. In other words, whether a learning model with desired performance can be constructed often depends on the learning dataset. Also, a method (so-called data poisoning) of attacking a model learned using a tampered learning dataset by maliciously tampering with a part of the learning dataset is known.
[0005] The present invention has been made in view of the above circumstances, and an object thereof is to provide a data management method capable of ensuring the reliability of learning data.
Means for Solving the Problems
[0006] A data management method according to one aspect of the present invention includes a management step in which management means associated with the storage manages proof information verifiable by a third party, which is issued to a storage that stores learning data satisfying predetermined conditions.
Brief Description of the Drawings
[0007] [Figure 1] This figure shows an example of a hardware configuration for realizing the data management method according to the embodiment. [Figure 2] This block diagram shows an example of the configuration of an information processing device according to the embodiment. [Figure 3] This is a conceptual diagram illustrating the concept of issuing certification information. [Modes for carrying out the invention]
[0008] Embodiments relating to the data management method will be described with reference to Figures 1 to 3. In Figure 1, a distributed network 1 comprising multiple computers may implement a first distributed ledger database. A distributed network 2 comprising multiple computers may implement a second distributed ledger database. At least one of the first distributed ledger database and the second distributed ledger database may be a blockchain.
[0009] The storage DS stores the training data LD used for machine learning of the model. The training data DS stored in the storage DS is data that satisfies certain conditions. These conditions may be at least one of the following: the data has been cleansed, it does not contain personal information, and it does not violate copyright law. An example of data cleansing when the training data DS is text data is to correct inconsistencies in spelling and typographical errors. Various existing methods can be applied to data cleansing. Therefore, a detailed explanation of data cleansing will be omitted.
[0010] The information processing device 10 manages the storage DS. The configuration of the information processing device 10 will be explained with reference to Figure 2. In Figure 2, the information processing device 10 comprises an arithmetic unit 11, a storage device 12, a communication device 13, an input device 14, and an output device 15. The arithmetic unit 11, storage device 12, communication device 13, input device 14, and output device 15 are connected via a data bus 16. The information processing device 10 may be a personal computer, a tablet terminal, or a smartphone.
[0011] The arithmetic unit 11 may have a processor. The arithmetic unit 11 may have a single processor or multiple processors. In other words, the arithmetic unit 11 may have one or more processors. Furthermore, the processor may be a multi-core processor. If the arithmetic unit 11 has a single processor that is a multi-core processor, then logically, the arithmetic unit 11 can be said to have multiple processors.
[0012] The processor may be at least one of the following: CPU (Central Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), and TPU (Tensor Processing Unit).
[0013] The storage device 12 may be at least one of the following: RAM (Random Access Memory), ROM (Read Only Memory), hard disk drive, magneto-optical disk drive, SSD (Solid State Drive), and optical disk array. In other words, the storage device 12 may be implemented by a single device or by multiple devices.
[0014] The communication device 13 may be capable of communicating with devices outside the information processing device 10. The communication device 13 may use either wired or wireless communication.
[0015] The input device 14 is a device capable of receiving information input to the information processing device 10 from an external source. The input device 14 may include an operating device (e.g., a keyboard, mouse, touch panel, etc.) that can be operated by the user of the information processing device 10. The input device 14 may include a recording medium reader capable of reading information recorded on a recording medium that can be attached to and detached from the information processing device 10, such as a USB (Universal Serial Bus) memory. When information is input to the information processing device 10 via the communication device 13 (in other words, when the information processing device 10 acquires information via the communication device 13), the communication device 13 may function as an input device.
[0016] The output device 15 is a device capable of outputting information to the outside of the information processing device 10. The output device 15 may have a display device capable of outputting visual information such as characters and images as the above information. The output device 15 may also have a speaker capable of outputting auditory information such as sound as the above information. The output device 15 may also have a vibration motor capable of outputting tactile information such as vibration as the above information. The output device 15 may also have a printer. The output device 15 may be capable of outputting information to a recording medium that can be attached to and detached from the information processing device 10, such as a USB memory stick. When the information processing device 10 outputs information via the communication device 13, the communication device 13 may function as an output device.
[0017] The storage device 12 is capable of storing desired data. The storage device 12 may store the computer program CP that the arithmetic unit 11 will execute. The storage device 12 may temporarily store data that the arithmetic unit 11 will use temporarily when the arithmetic unit 11 is executing the computer program CP.
[0018] Furthermore, the computer program CP may be recorded on a non-temporary recording medium that is readable by a computer. In this case, the computer program CP may be stored in the storage device 12 by reading the recording medium using a recording medium reading device (not shown) provided by the information processing device 10. Furthermore, at least one of the following may be used as the recording medium: an optical disc, a magnetic medium, a magneto-optical disc, a semiconductor memory, and any other medium capable of storing a program. Furthermore, the computer program CP may be obtained from an external device (not shown) of the information processing device 10 via a communication device 13. In other words, the computer program CP may be downloaded from an external device to the storage device 12 of the information processing device 10.
[0019] The arithmetic unit 11 (for example, a processor) may execute the processing that the information processing device 10 should perform together with the memory device 12 in which the computer program CP is stored (in other words, together with the memory device 12 and the computer program CP stored in the memory device 12). For example, by the arithmetic unit 11 executing the computer program CP, a logical functional block for executing the processing that the information processing device 10 should perform may be realized within the arithmetic unit 11 (for example, within the processor).
[0020] For example, the arithmetic unit 11 of the information processing device 10 may calculate the hash value of the training data LD stored in the storage DS. The information processing device 10 may register the calculated hash value in a second distributed ledger database via the distributed network 2. As a result, the hash values related to the training data LD may be managed by the second distributed ledger database.
[0021] After the hash value related to the learning data LD stored in the storage DS is registered in the second distributed ledger database, the information processing device 10 may request the issuer V (see FIG. 3) of the provable proof information verifiable by a third party to issue the proof information related to the storage DS. Note that, as an example of the provable proof information verifiable by a third party, Verifiable Credentials (VC) can be cited.
[0022] The issuer V that has received the proof information issuance request from the information processing device 10 may examine whether the storage DS satisfies the predetermined issuance requirements. For example, the predetermined issuance requirements may include that the learning data LD satisfies the predetermined requirements and that the hash value related to the learning data LD is registered in the distributed ledger database.
[0023] If the storage DS satisfies the predetermined issuance conditions, the issuer V may issue the proof information related to the storage DS. The issued proof information may be transmitted to the information processing device 10 that manages the storage DS (in other words, is associated with the storage DS). Here, the proof information may include issuer identification information for identifying the issuer V, storage identification information for identifying the storage DS, and a proof such as a digital signature. When the issuer V issues the proof information, the issuer V may register the issuer identification information, the storage identification information, the public key related to the issuer V, etc. in the first distributed ledger database via the distributed network 1.
[0024] The information processing device 10 that has received the issued proof information stores the proof information in the management means (for example, a wallet) associated with the storage identification information. Note that at least a part of the management means may be realized by the storage device 12.
[0025] Note that the first distributed ledger database may be referred to as "Verifiable Data Registry". At least one of the issuer identification information and the storage identification information may be a decentralized identifier (DID).
[0026] An example of a method for verifying certification information related to a storage DS is described below. Based on the issuer identification information contained in the certification information related to the storage DS, the public key of issuer V may be obtained from the first distributed ledger database. Subsequently, the digital signature contained in the certification information related to the storage DS may be verified using the public key of issuer V.
[0027] (Technical effects) As described above, the hash value of the training data LD stored in the storage DS is managed by the second distributed ledger database. If at least a portion of the training data LD is tampered with, the hash value of the tampered training data LD will differ from the hash value of the training data LD registered in the second distributed ledger database. Therefore, by managing the hash value of the training data LD in the second distributed ledger database, it is possible to suppress tampering with the training data LD stored in the storage DS. In addition, a third-party verifiable certification (e.g., VC) is issued to the storage DS. This certification can prove the legitimacy of the storage DS. As a result, the reliability of the training data LD stored in the storage DS can be guaranteed.
[0028] Various aspects of the invention derived from the embodiments described above are described below.
[0029] A data management method according to one aspect of the invention includes a management step in which a management means associated with a storage device manages a third-party verifiable certification information issued to a storage device that stores learning data satisfying predetermined conditions.
[0030] In the data management method according to the above embodiment, the hash value related to the training data may be registered on the blockchain. This configuration makes it possible to suppress tampering with the training data.
[0031] In the data management method according to the above embodiment, the predetermined conditions may be at least one of the following: the data has been cleansed, it does not contain personal information, and it does not violate copyright law. With this configuration, high-quality training data can be provided.
[0032] In the data management method according to the above embodiment, the certification information may be verifiable credentials.
[0033] 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. Data management methods involving such modifications are also included within the technical scope of the present invention. [Explanation of symbols]
[0034] 1, 2…Distributed network, 10…Information processing device, DS…Storage
Claims
1. This includes a management process in which a management means associated with a storage device manages third-party verifiable certification information issued to a storage device that stores learning data that meets predetermined conditions. Data management methods.
2. The hash value related to the aforementioned learning data is registered on the blockchain. The data management method according to claim 1.
3. The aforementioned conditions are at least one of the following: the data has been cleansed, it does not contain personal information, and it does not violate copyright law. The data management method according to claim 1.
4. The aforementioned certification information is Verified Credentials. The data management method according to claim 1.