Computer system, and data transmission method

Abstract

To increase the number of data for use in training of a machine learning model, while securing security of information.SOLUTION: A computer system according to the present invention comprises: a collection system for collecting business data including values of a plurality of items from a plurality of business systems; and a plurality of data processing systems. The collection system is coupled to an AI processing system configured to execute, through use of encrypted business data, training processing of generating a machine learning model. Each of the plurality of data processing systems acquires the business data from one of the business systems; encrypts a value of any one of the plurality of items of the business data through use of an irreversible encryption algorithm to generate the encrypted business data; and transmits the encrypted business data to the collection system. The collection system transmits the encrypted business data to the AI processing system.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to a technique for collecting data used to train machine learning models. [Background technology]

[0002] In recent years, labor shortages have become a problem in various industries, and the use of AI in business operations has attracted attention. AI that supports business operations, i.e., generating machine learning models, requires a large amount of training data.

[0003] When there is little training data, the inference results of the machine learning model are unreliable, and workers must review and correct the inference results, thereby negating the benefits of using AI.

[0004] A known technique for increasing training data is described in, for example, Patent Document 1. Patent Document 1 discloses a "training data creation system including: an acquisition unit that acquires a first image, a second image, first correct answer information corresponding to the first image, and second correct answer information corresponding to the second image; a first neural network that generates a first feature map by inputting the first image and a second feature map by inputting the second image; a feature map synthesis unit that generates a synthesized feature map by replacing a portion of the first feature map with a portion of the second feature map; a second neural network that generates output information based on the synthesized feature map; an output error calculation unit that calculates an output error based on the output information, the first correct answer information, and the second correct answer information; and a neural network update unit that updates the first neural network and the second neural network based on the output error." The technique described in Patent Document 1 is effective for machine learning models that perform image recognition. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] International Publication No. 2021 / 176605 Summary of the Invention [Problem to be solved by the invention]

[0006] One way to increase the amount of training data for machine learning models that support business operations is to collect training data from users (companies) in the same industry. Even if the amount of training data that can be obtained from a single user is small, it is possible to collect a large amount of training data by collecting training data from users in the same industry.

[0007] However, since the business data provided as training data contains confidential information, providing it to an external party as is poses security issues. [Means for solving the problem]

[0008] A representative example of the invention disclosed in the present application is as follows: A computer system including a collection system that collects business data including values ​​of multiple items from multiple business systems, and multiple data processing systems that process the business data, wherein the collection system is connected to an AI processing system that executes at least one of a learning process that uses encrypted business data to generate a machine learning model and an inference process that performs inference using the machine learning model, the data processing system acquires the business data from one of the business systems, encrypts values ​​of any of the items in the business data using an irreversible encryption algorithm to generate the encrypted business data, transmits the encrypted business data to the collection system, and the collection system transmits the encrypted business data to the AI ​​processing system. [Effects of the Invention]

[0009] According to the present invention, it is possible to increase the amount of data used for training a machine learning model while ensuring information security. Problems, configurations, and effects other than those described above will become clear from the following description of the embodiments. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a computer system according to a first embodiment. [Figure 2] FIG. 2 illustrates an example of a configuration of a data processing server according to the first embodiment. [Figure 3] FIG. 10 is a sequence diagram illustrating the flow of a process for generating reshaping rule information in the computer system of the first embodiment. [Figure 4A] FIG. 2 is a diagram illustrating an example of a user interface presented by the data processing server according to the first embodiment. [Figure 4B] FIG. 2 is a diagram illustrating an example of a user interface presented by the data processing server according to the first embodiment. [Figure 4C] FIG. 2 is a diagram illustrating an example of a user interface presented by the data processing server according to the first embodiment. [Figure 5] FIG. 10 is a sequence diagram illustrating the flow of a process for transmitting business data in the computer system according to the first embodiment. [Figure 6A] FIG. 10 is a diagram illustrating a specific example of processing by the data processing server according to the first embodiment. [Figure 6B] FIG. 10 is a diagram illustrating a specific example of processing by the data processing server according to the first embodiment. [Figure 6C] FIG. 10 is a diagram illustrating a specific example of processing by the data processing server according to the first embodiment. [Figure 7] FIG. 10 is a sequence diagram illustrating the flow of a process for transmitting business data in the computer system according to the first embodiment. [Figure 8A] FIG. 10 is a diagram illustrating a specific example of processing by the data processing server according to the first embodiment. [Figure 8B] FIG. 10 is a diagram illustrating a specific example of processing by the data processing server according to the first embodiment. [Figure 8C] FIG. 10 is a diagram illustrating a specific example of processing by the data processing server according to the first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present invention will be described with reference to the drawings. However, the present invention should not be construed as being limited to the description of the embodiments shown below. Those skilled in the art will readily understand that the specific configuration can be changed without departing from the concept or spirit of the present invention.

[0012] In the configuration of the invention described below, the same or similar configurations or functions are denoted by the same reference numerals, and redundant explanations will be omitted.

[0013] In this specification, the terms "first," "second," "third," etc. are used to identify components and do not necessarily limit the number or order. [Example]

[0014] FIG. 1 is a diagram illustrating an example of the configuration of a computer system according to a first embodiment.

[0015] The computer system is composed of a data collection server 100, an AI processing server 101, and multiple business systems 102. The data collection server 100 is connected to the multiple business systems 102 via a network such as a LAN (Local Area Network). The data collection server 100 is also connected to the AI ​​processing server 101 directly or via a network.

[0016] The business system 102 includes a business server 110, a data processing server 111, and an adapter 112. The business system 102 may be an on-premise system or a cloud system such as SaaS (Software as a Service).

[0017] The business server 110 executes various processes related to the business. The data processing server 111 acquires business data including values ​​of multiple items from the business server 110 and executes various processes on the business data. The adapter 112 communicates with the data collection server 100. The adapter 112 may execute analysis, processing, communication processing, etc. of the business data as needed.

[0018] The data collection server 100 collects business data from a business system 102 and transmits it to the AI ​​processing server 101. The AI ​​processing server 101 performs learning of a machine learning model and inference using the machine learning model. For example, a machine learning model may be one that supports port operations that handle the import and export of container cargo.

[0019] The AI ​​processing server 101 may have the functions of the data collection server 100. The data collection server 100 and the AI ​​processing server 101 may be constructed on independent computer systems (for example, SaaS), or may be constructed on the same computer system.

[0020] FIG. 2 is a diagram illustrating an example of the configuration of the data processing server 111 according to the first embodiment.

[0021] The data processing server 111 includes a processor 200, a memory 201, and a network interface 202. The data processing server 111 may also include storage devices such as a hard disk drive (HDD) and a solid state drive (SSD), input devices such as a keyboard, a mouse, and a touch panel, and an output device such as a display.

[0022] The processor 200 executes a program stored in the memory 201. The processor 200 executes processing in accordance with the program, thereby operating as a functional unit (module) that realizes a specific function. In the following description, when a processing is described using a functional unit as the subject, it indicates that the processor 200 is executing a program that realizes the functional unit.

[0023] The memory 201 stores programs executed by the processor 200 and information used by the programs. The memory 201 is also used as a work area. In the first embodiment, the memory 201 stores programs that realize the shaping unit 210 and the encryption unit 211, and also stores shaping rule information 220.

[0024] The shaping unit 210 performs shaping processing on the business data. In the shaping processing of the business data, the names and values ​​of the items in the business data are converted based on the shaping rule information 220.

[0025] The encryption unit 211 encrypts business data. Specifically, the encryption unit 211 encrypts the business data using a hash function. The same hash function is set in the data processing server 111 of each business system 102. This means that the same value becomes the same hash value, making it possible to integrate and consolidate the encrypted business data of each business system 102.

[0026] It should be noted that the functional units of the data processing server 111 may be configured such that multiple functional units are combined into one functional unit, or one functional unit may be divided into multiple functional units for each function.

[0027] The formatting rule information 220 stores formatting rule data for item names and formatting rule data for item values. The formatting rule data for item names is composed of the name of the item before conversion and the name of the item after conversion. The formatting rule data for item values ​​is composed of the item name, unit system, and converted unit.

[0028] The shaping rule information 220 may be set in advance or may be generated based on user input. Here, a method for generating the shaping rule information 220 based on user input will be described.

[0029] Fig. 3 is a sequence diagram illustrating the flow of a process for generating the reshaping rule information 220 in the computer system of the embodiment 1. Fig. 4A, Fig. 4B, and Fig. 4C are diagrams illustrating an example of a user interface presented by the data processing server 111 of the embodiment 1.

[0030] The data processing server 111 presents the user interface 400 (S101).

[0031] The data processing server 111 first presents a user interface 400 shown in Fig. 4A. The user interface 400 includes buttons 401, 402, 403, and 404. The button 404 is an operation button for logging out.

[0032] When button 401 is operated, user interface 400 transitions to the state shown in Fig. 4B. User interface 400 displays a setting table 410 and buttons 420 and 421. Setting table 410 displays entries made up of item names (before conversion) 411 and item names (after conversion) 412. Setting table 410 has entries equal to the number of items included in the business data.

[0033] Item name (before conversion) 411 is a field for storing the name of an item of business data. Item name (after conversion) 412 is a field for setting the name of the item after conversion. In item name (after conversion) 412, the names of items common to the computer system are displayed, for example, in a pull-down format.

[0034] After setting an appropriate name in item name (after conversion) 412 for each item, the user operates button 420 to set setting table 410 as item information. After button 420 is operated, user interface 400 transitions to the state shown in Fig. 4A. Note that when button 421 is operated, item information is not set and user interface 400 transitions to the state shown in Fig. 4A.

[0035] When button 402 is operated, user interface 400 transitions to the state shown in Fig. 4C. User interface 400 displays a setting table 430 and buttons 440 and 441. Setting table 430 displays entries each consisting of an item name (before conversion) 431, SI unit system 432, and data unit 433. Setting table 430 has as many entries as there are items for which values ​​with units are stored.

[0036] Item name (before conversion) 431 is a field for storing the name of an item in which a value with a unit is stored. SI unit system 432 is a field for setting the type of unit for the value of the item. In SI unit system 432, the names of units common to computer systems are displayed, for example, in a pull-down format. Data unit 433 is a field for setting the unit of the value stored in the item. In data unit 433, the unit of the value is displayed, for example, in a pull-down format.

[0037] After setting appropriate values ​​for SI unit system 432 and data unit 433, the user operates button 440 to set setting table 430 as unit information. After button 440 is operated, user interface 400 transitions to the state shown in Fig. 4A. Note that when button 441 is operated, unit information is not set and user interface 400 transitions to the state shown in Fig. 4A.

[0038] When the button 403 is operated, the data processing server 111 transmits the item information and the unit information to the data collection server 100 (S102). Note that only either the item information or the unit information may be transmitted to the data collection server 100.

[0039] When the data collection server 100 receives the item information and unit information, it generates the shaping rule information 220 (S103). Specifically, the following process is executed.

[0040] (S103-1) The data collection server 100 generates each entry of the item information as formatting rule data for the name of the item.

[0041] (S103-2) The data collection server 100 associates the item name and unit with a unit common to the computer system for each entry of the unit information, and generates formatting rule data for the value of the item.

[0042] The data collection server 100 transmits the shaping rule information 220 to the data processing server 111 (S104).

[0043] Fig. 5 is a sequence diagram illustrating the flow of a process for transmitting business data in the computer system of the embodiment 1. Fig. 6A, Fig. 6B, and Fig. 6C are diagrams illustrating a specific example of the process of the data processing server 111 of the embodiment 1. Fig. 5 illustrates the process for transmitting business data used as learning data.

[0044] The business server 110 transmits business data for learning to the data processing server 111 (S201). Note that the data processing server 111 may acquire the business data from the business server 110.

[0045] The data processing server 111 performs formatting processing on the received business data based on the formatting rule information 220 (S202).

[0046] Specifically, the shaping unit 210 converts the names of the business data items based on the shaping rule data for the item names. Also, the shaping unit 210 converts the values ​​of the items into values ​​in the specified units based on the shaping rule data for the item values.

[0047] For example, when a table 600 including multiple pieces of business data as shown in Fig. 6A is received, the rectifying unit 210 converts the item names and item values ​​as shown in Fig. 6B. In Fig. 6B, the "delivery date and time" is converted to "unloading date and time", the "container No." is converted to "container ID", the "shipper name" is converted to "shipper", the "load" is converted to "cargo", and the "load weight" is converted to "weight". In addition, the value of "load weight" is converted from "kg" to "t".

[0048] The data processing server 111 performs encryption processing on the business data that has been subjected to the shaping processing (S203).

[0049] Specifically, the encryption unit 211 inputs the value of a predetermined item into a hash function to calculate a hash value of the value of the item. For example, as shown in Fig. 6C, the values ​​of "container ID," "shipper," and "cargo" are encrypted.

[0050] It is assumed that the items to be encrypted are set in advance. Even if a value indicating a type such as a product type or a gender is encrypted, the difference between types can be distinguished. Therefore, in the first embodiment, the items to be encrypted are set based on the impact of loss of information necessary for inference of the machine learning model and on ensuring security.

[0051] The data processing server 111 transmits the encrypted business data (encrypted business data) to the data collection server 100 via the adapter 112 (S204, S205).

[0052] When the data collection server 100 receives encrypted business data as learning data, it transmits a reception response to the adapter 112 (S206), and also transmits the encrypted business data to the AI ​​processing server 101 (S207).

[0053] The AI ​​processing server 101 executes a learning process using the encrypted business data received as learning data (S208). The learning process is executed at any timing. By acquiring encrypted business data from multiple business systems 102, the number of learning data samples can be increased. On the other hand, the encrypted business data can be handled uniformly because it has undergone a formatting process. Furthermore, the encrypted business data can ensure security because the values ​​of items that need to be kept confidential are encrypted.

[0054] Fig. 7 is a sequence diagram illustrating the flow of business data transmission processing in the computer system of Example 1. Fig. 8A, Fig. 8B, and Fig. 8C are diagrams illustrating specific examples of processing by the data processing server 111 of Example 1. Fig. 7 illustrates the transmission processing of business data used as data for inference.

[0055] The business server 110 transmits business data for inference to the data processing server 111 (S301). Note that the data processing server 111 may acquire the business data from the business server 110.

[0056] The data processing server 111 performs formatting processing on the received business data (S302) based on the formatting rule information 220. The processing in S302 is the same as the processing in S202.

[0057] For example, when business data 800 as shown in Fig. 8A is received, the rectifying unit 210 converts the item names and item values ​​as shown in Fig. 8B. In Fig. 8B, the "delivery date and time" is converted to the "unloading date and time", the "container No." is converted to the "container ID", the "shipper name" is converted to the "shipper", the "load" is converted to the "cargo", and the "load weight" is converted to the "weight". In addition, the value of the "load weight" is converted from kg to t.

[0058] The data processing server 111 performs encryption processing on the business data that has been subjected to the shaping processing (S303). The processing in S303 is the same as the processing in S203.

[0059] For example, as shown in FIG. 8C, the values ​​of "Container ID," "Shipper," and "Cargo" are encrypted.

[0060] The data processing server 111 transmits the encrypted business data (encrypted business data) to the data collection server 100 via the adapter 112 (S304, S305).

[0061] When the data collection server 100 receives encrypted business data as data for inference, it sends a reception response to the adapter 112 (S306) and also sends the encrypted business data to the AI ​​processing server 101 (S307).

[0062] The AI ​​processing server 101 executes inference processing by inputting the encrypted business data received as data for inference into the machine learning model (S308). The encrypted business data can be handled uniformly because it has been formatted. Furthermore, security can be ensured because the values ​​of items that need to be kept confidential are encrypted in the encrypted business data.

[0063] The AI ​​processing server 101 transmits the results of the inference processing to the business server 110 via the adapter 112 (S309, S310).

[0064] Although the data processing server 111 has been described as being configured independently of the business server 110 and the adapter 112, the present invention is not limited to this. For example, the business server 110 or the adapter 112 may have the functions of the data processing server 111.

[0065] The present invention is not limited to the above-described embodiments, but includes various modifications. For example, the above-described embodiments are provided to explain the present invention in detail, and the present invention is not necessarily limited to those including all of the described configurations. Furthermore, some of the configurations of each embodiment can be added to, deleted from, or replaced with other configurations.

[0066] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The present invention can also be realized by software program code that implements the functions of the embodiments. In this case, a storage medium on which the program code is recorded is provided to a computer, and a processor included in the computer reads the program code stored in the storage medium. In this case, the program code itself read from the storage medium implements the functions of the above-described embodiments, and the program code itself and the storage medium on which it is stored constitute the present invention. Examples of storage media for providing such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, solid-state drives (SSDs), optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, and ROMs.

[0067] Furthermore, the program code that realizes the functions described in this embodiment can be implemented in a wide range of program or script languages, such as assembler, C / C++, perl, Shell, PHP, Python, and Java (registered trademark).

[0068] Furthermore, the program code of the software that realizes the functions of the embodiments may be distributed via a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the processor of the computer may read and execute the program code stored in the storage means or storage medium.

[0069] In the above-described embodiment, the control lines and information lines are those that are considered necessary for the explanation, and not all control lines and information lines are necessarily shown in the product. All components may be interconnected. [Explanation of symbols]

[0070] 100 Data Collection Servers 101 AI processing server 102 Business Systems 110 Business Server 111 Data Processing Server 112 Adapter 200 processors 201 Memory 202 Network Interface 210 Plastic Surgery Department 211 Encryption section 220 Formatting rule information 400 User Interface

Claims

1. A computer system comprising: a collection system that collects business data including values ​​of a plurality of items from a plurality of business systems; and a plurality of data processing systems that process the business data, The collection system is connected to an AI processing system that uses encrypted business data to execute at least one of a learning process that generates a machine learning model and an inference process that performs inference using the machine learning model; The data processing system includes: Acquire the business data from one of the business systems; generating the encrypted business data by encrypting the value of any of the items of the business data using an irreversible encryption algorithm; transmitting the encrypted business data to the collection system; A computer system characterized in that the collection system transmits the encrypted business data to the AI ​​processing system.

2. 2. The computer system of claim 1, the data processing system encrypts a value of any of the items of the business data using a hash function; A computer system characterized in that the hash functions used by the plurality of data processing systems are the same.

3. 3. The computer system according to claim 2, The data processing system includes: execute a formatting process in which a process of converting the name of at least one of the items into a name of an item common to the plurality of business systems and a process of converting the value of at least one of the items into a value in a unit common to the plurality of business systems are executed; A computer system characterized in that the value of any of the items of the business data on which the shaping process has been performed is encrypted.

4. 4. The computer system according to claim 3, the data processing system acquires setting information relating to the names of the items to be subjected to the shaping process and the values ​​of the items, and transmits the setting information to the collection system; The collection system comprises: generating shaping rule information by determining a shaping rule for the name of the item to be shaped and a shaping rule for the value of the item to be shaped based on the setting information; transmitting the shaping rule information to the data processing system; The data processing system executes the formatting process based on the formatting rule information.

5. A data transmission method executed by a computer system, comprising: the computer system includes a collection system that collects business data including values ​​of a plurality of items from a plurality of business systems, and a plurality of data processing systems that process the business data; The collection system is connected to an AI processing system that uses encrypted business data to execute at least one of a learning process that generates a machine learning model and an inference process that performs inference using the machine learning model; The data transmission method includes: a first step in which the data processing system acquires the business data from one of the business systems; a second step in which the data processing system generates the encrypted business data by encrypting a value of any of the items of the business data using an irreversible encryption algorithm; a third step in which the data processing system transmits the encrypted transaction data to the collection system; A data transmission method comprising: a fourth step in which the collection system transmits the encrypted business data to the AI ​​processing system.

6. 6. The data transmission method according to claim 5, the second step includes a step in which the data processing system encrypts a value of any of the items of the business data using a hash function; A data transmission method, wherein the hash functions used by the plurality of data processing systems are the same.

7. 7. The data transmission method according to claim 6, The second step includes: a fifth step in which the data processing system executes a formatting process in which a process of converting the name of at least one of the items into a name of an item common to the plurality of business systems and a process of converting the value of at least one of the items into a value in a unit common to the plurality of business systems are executed; a sixth step in which the data processing system encrypts the value of any of the items of the business data on which the formatting process has been performed.

8. 8. The data transmission method according to claim 7, The data processing system acquires setting information relating to the names of the items to be subjected to the shaping process and the values ​​of the items, and transmits the setting information to the collection system; The collection system generates shaping rule information by determining a shaping rule for the name of the item to be shaped based on the setting information and a shaping rule for the value of the item to be shaped based on the setting information; the collection system transmitting the shaping rule information to the data processing system; The data transmission method, wherein the fifth step includes a step in which the data processing system executes the shaping process based on the shaping rule information.

Images