Data control method and related device and medium program product

Abstract

The embodiment of the present application provides a data management method and related devices and medium program products, the method comprising: determining the target learning task and the first training set corresponding to the current working condition information according to the current working condition information and the reference working condition information pre-stored in the learning training set database; removing the first training data set with problems from the first training set to obtain the second training set; training according to the second training set and the target learning task to obtain the target learning result; storing the target learning result in the learning training set database to obtain the updated learning training set database. This can not only improve the efficiency of machine learning, but also improve the accuracy of machine learning.

Description

Technical Field

[0001] The present application belongs to the general data processing technology field of the Internet industry, and specifically relates to a data management method and related devices and medium program products. Background Art

[0002] At present, the traditional machine learning method creates a new training set each time, does not make full use of the existing historical training data, resulting in low machine learning efficiency, and does not eliminate training data that causes quality problems in production batches, resulting in poor machine learning accuracy. Summary of the invention

[0003] The present application provides a data management method and related devices and medium program products, in order to improve the efficiency and accuracy of machine learning.

[0004] In a first aspect, an embodiment of the present application provides a data management method, which is applied to a server and includes:

[0005] Determine a target learning task and a first training set corresponding to the current operating condition information according to the current operating condition information and the reference operating condition information pre-stored in the learning training set database, wherein the target learning task is used to indicate learning of production parameters under the current operating condition through a preset model, and the first training set is used to indicate a training data set used for learning in the target learning task;

[0006] Remove the problematic first training data set from the first training set to obtain a second training set;

[0007] Perform training according to the second training set and the target learning task to obtain a target learning result;

[0008] The target learning result is stored in the learning training set database to obtain an updated learning training set database.

[0009] In a second aspect, an embodiment of the present application provides a data control device, which is applied to a server and includes:

[0010] A determination unit is used to determine a target learning task and a first training set corresponding to the current operating condition information based on the current operating condition information and the reference operating condition information pre-stored in the learning training set database, wherein the target learning task is used to indicate learning of production parameters under the current operating condition through a preset model, and the first training set is used to indicate a training data set used for learning in the target learning task; a clearing unit is used to clear a problematic first training data set from the first training set to obtain a second training set; a training unit is used to perform training based on the second training set and the target learning task to obtain a target learning result; a storage unit is used to store the target learning result in the learning training set database to obtain an updated learning training set database.

[0011] In a third aspect, an embodiment of the present application provides a server, comprising a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the program includes instructions for executing the steps in the first aspect of the embodiment of the present application.

[0012] In a fourth aspect, an embodiment of the present application provides a computer storage medium for storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute some or all of the steps described in the first aspect of the embodiment.

[0013] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps described in the first aspect of the embodiment of the present application. The computer program product may be a software installation package.

[0014] It can be seen that in the embodiment of the present application, the target learning task and the first training set corresponding to the current working condition information are first determined based on the current working condition information and the reference working condition information pre-stored in the learning training set database, and then the problematic first training data set is removed from the first training set to obtain the second training set, and then training is performed based on the second training set and the target learning task to obtain the target learning result, and finally the target learning result is stored in the learning training set database to obtain an updated learning training set database. In this way, each training data is stored in the database for subsequent reuse in machine learning, which can effectively improve the efficiency of machine learning, and remove abnormal training data to ensure the accuracy of machine learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a schematic diagram of a system architecture provided by an embodiment of the present application;

[0017] Figure 2 It is a flowchart of a data control method provided in an embodiment of the present application;

[0018] Figure 3 This is a schematic diagram of the first client interaction page provided in an embodiment of the present application;

[0019] Figure 4 This is a schematic diagram of a second client interaction page provided in an embodiment of the present application;

[0020] Figure 5 This is a schematic diagram of the third client interaction page provided in the embodiment of the present application;

[0021] Figure 6 This is a schematic diagram of the fourth client interaction page provided in an embodiment of the present application;

[0022] Figure 7 is a schematic diagram of clearing abnormal training data provided by an embodiment of the present application;

[0023] Figure 8a This is a block diagram of the functional units of a data control device provided in an embodiment of the present application;

[0024] Figure 8b It is a block diagram of the functional units of another data control device provided in an embodiment of the present application;

[0025] Fig. 9 It is a structural block diagram of a server provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0027] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.

[0028] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0029] The following is an introduction to the relevant terms involved in this application.

[0030] Production batch: refers to a production process in which raw materials are processed by production equipment to obtain production results. Correspondingly, each production batch has a corresponding production batch number to distinguish different production batches.

[0031] Working condition: refers to the combination of the change range of the characteristic values ​​of the input factors of the production process. In actual production, the combination of these factors will be divided into zones and coded according to the characteristic range of the factors to form different working condition codes. Different working condition combinations will have a significant impact on the control parameters of the production process.

[0032] Machine learning: Machine learning evolved into a branch of artificial intelligence in the second half of the 20th century. It uses self-learning algorithms to obtain knowledge from data to make predictions. Machine learning does not require manual analysis of large amounts of data in advance, and then extracting rules and building models. Instead, it provides a more effective way to capture knowledge in the data and gradually improve the performance of the prediction model to complete data-driven decision-making.

[0033] At present, when traditional machine learning training sets are managed, a new training set is created each time, which does not make full use of the existing historical training data, resulting in low machine learning efficiency. In addition, training data that causes quality problems in production batches is not eliminated, resulting in poor machine learning accuracy.

[0034] To solve the above problems, an embodiment of the present application provides a data management method, which can be applied to the field of manufacturing business. This method can be used to determine learning tasks and training sets based on reference working condition information stored in a database when an enterprise needs to perform machine learning in the face of different working conditions, and to clear abnormal training data in the training set, thereby improving the efficiency of machine learning while ensuring the accuracy of machine learning. The present application can be applied to a variety of application scenarios that require machine learning training set management, including but not limited to the application scenarios mentioned above.

[0035] The following introduces the system architecture involved in the embodiments of the present application.

[0036] See also Figure 1 , Figure 1 Schematic diagram of a system architecture provided by an embodiment of the present application. Figure 1 As shown, the system architecture 10 includes a server 11 and a client 12, wherein the server 11 is in communication connection with the client 12, wherein when the server 11 determines the target learning task and the first training set corresponding to the current working condition information according to the current working condition information and the reference working condition information pre-stored in the learning training set database, the server 11 will send a corresponding message to the client 12 according to different detection results to instruct the client 12 to send different instructions, such as task creation instructions, task merging instructions or training data use confirmation instructions, etc., so as to obtain the target learning task and the first training set according to these instructions. Among them, the server 11 can be a server, or a server cluster composed of several servers, or a cloud computing service center, and the client 12 can be a mobile terminal, a tablet computer, a laptop computer, a vehicle-mounted terminal, etc.

[0037] The following describes a data management method provided in an embodiment of the present application.

[0038] See also Figure 2 , Figure 2 is a flow chart of a data control method provided in an embodiment of the present application, wherein the method is applied to a server, such as Figure 2 As shown, the data control method includes:

[0039] Step 201 : determining a target learning task and a first training set corresponding to the current working condition information according to the current working condition information and reference working condition information pre-stored in a learning training set database.

[0040] Among them, the target learning task is used to indicate the learning of production parameters under the current working condition through a preset model, and the first training set is used to indicate the training data set used for learning in the target learning task. The current working condition information refers to the new working condition that needs machine learning in the current production batch. The information of the new working condition is the current working condition information. Exemplarily, the information may include the working condition type, working condition number, etc. If the current working condition information is an old working condition, the current production mode of the old working condition will be further determined, and different production strategies will be implemented according to different production modes, and learning tasks will not be created or merged. Therefore, the current working condition information in this application refers to the new working condition that needs machine learning. The learning training set database stores training records of historical working conditions. Each historical working condition has a unique working condition number to distinguish different working conditions. Each working condition may have only one learning task, multiple learning tasks, or no corresponding learning task but corresponding historical training data.

[0041] Step 202: remove the problematic first training data set from the first training set to obtain a second training set.

[0042] Among them, the quality problems include: problems with the production equipment parameters (i.e., process parameters) of a production batch or abnormalities in the quality inspection of production materials, all of which will lead to unqualified quality of the final product. The current operating condition information is the operating condition information that appears under the target production batch. Therefore, when deploying learning tasks for the operating condition information, it is necessary to remove the training data in the training set that will cause quality problems in the target production batch, obtain a training set without abnormal data, and then use the training set without abnormal data for machine learning, thereby ensuring the accuracy of machine learning.

[0043] Step 203: Perform training according to the second training set and the target learning task to obtain a target learning result.

[0044] Among them, the target learning result is used to characterize that the current working condition information under the target production batch has completed machine learning, and the target learning result includes the production parameters corresponding to the current working condition information, and the production parameters can be parameters of production equipment or component parameters, and the component parameters can include raw material parameters and intermediate product parameters. At the same time, the current working condition information is associated with the working condition code and the second training set, and is marked by the server, and the training data in the second training set is recorded as the historical training data of the current working condition, and the target learning task is recorded as the historical learning task of the current working condition. The target learning result includes the above-mentioned recording results.

[0045] Step 204, storing the target learning result in the learning training set database to obtain an updated learning training set database.

[0046] Among them, after the target learning results are stored in the learning training set database, the training data and learning tasks therein can be used for machine learning tasks when encountering new working conditions in the future production process, avoiding the need to create a new training set every time the training set is used for machine learning, which reduces the efficiency of machine learning.

[0047] It can be seen that in this example, the target learning task and the first training set corresponding to the current working condition information are first determined based on the current working condition information and the reference working condition information pre-stored in the learning training set database, and then the first training data set that causes the target production batch corresponding to the current working condition information to have quality problems is removed from the first training set to obtain the second training set, and then training is performed based on the second training set and the target learning task to obtain the target learning result, and finally the target learning result is stored in the learning training set database to obtain the updated learning training set database. In this way, each training data is stored in the database for subsequent reuse in machine learning, which can effectively improve the efficiency of machine learning, and remove abnormal training data to ensure the accuracy of machine learning.

[0048] In a possible example, determining a target learning task and a first training set corresponding to the current operating condition information based on the current operating condition information and reference operating condition information pre-stored in a learning training set database includes: detecting that there is no historical learning task corresponding to the current operating condition information and no historical training data corresponding to the current operating condition information in the reference operating condition information; obtaining an original training set with X training data; receiving a task creation instruction from a client; creating the target learning task according to the task creation instruction; and determining the first training set according to the task creation instruction and the original training set with X training data, wherein the value of X is greater than a target preset value.

[0049] Among them, the server detects that there is no corresponding historical learning task and corresponding historical training data in the learning training set database, indicating that there is no available historical training data. In this case, the server needs to create a new learning task and obtain the original training set for machine learning. Figure 3 , Figure 3 This is a schematic diagram of the first client interaction page provided in the embodiment of the present application. The server sends a message to the client, and a pop-up window such as Figure 3In the interactive page shown, the user clicks the "Create Task" control to cause the client to send the task creation instruction to the server. Since there is no available historical training data in the database, the original training set obtained is the first training set, and to ensure the learning effect, the number of training data in the first training set should be greater than the target preset value, and the size of the target preset value is flexibly determined according to different working conditions.

[0050] It can be seen that in this example, the server detects that there is no available historical data in the database, and then creates a new learning task and obtains the original training set as the first training set. It also sets the minimum number of training data in the first training set to ensure the effect of machine learning and avoid deviations in learning effects due to insufficient training samples resulting from a small number of training data.

[0051] In a possible example, determining a target learning task and a first training set corresponding to the current operating condition information based on the current operating condition information and reference operating condition information pre-stored in a learning training set database includes: detecting that there are multiple historical learning tasks corresponding to the current operating condition information and historical training sets corresponding to the historical learning tasks in the reference operating condition information; obtaining an original training set with X training data; receiving a task merging instruction and a task selection instruction from a client; selecting N historical learning tasks through the task selection instruction; merging the N historical learning tasks through the task merging instruction to obtain the target learning task; merging the original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks through the task merging instruction to obtain the first training set, wherein the value of X+Y is greater than the target preset value.

[0052] Among them, the server detects that there are already multiple historical learning tasks corresponding to the current operating condition information in the learning training set database. At this time, in order to improve the efficiency of machine learning, the multiple historical learning tasks will be merged as the target learning task to merge the historical training sets in the multiple historical learning tasks. Since the merged historical training set may have the situation that the number of training data in the training set does not meet the standard, or some training data is unavailable, the original training set will still be obtained first when merging learning tasks. At this time, the X training data in the original training set can be adjusted according to the size of the number Y of historical training data formed after the merger, as long as the number of training data in the first training set (X+Y) is greater than the target preset value. For example, please refer to Figure 4 , Figure 4 This is a schematic diagram of a second client interaction page provided in an embodiment of the present application. The server sends a message to the client, and a pop-up window such as Figure 4The interactive interface shown in the figure displays all the historical learning tasks that can be merged. The user sends the task selection instruction to the server by checking the checkbox in front of the historical learning task, and then clicks the "Merge Task" control to make the client send the task merging instruction to the server. The server merges the selected historical learning tasks according to the task selection instruction and the task merging instruction to obtain the target learning task, and merges the original training set with the historical training set to obtain the first training set with (X+Y) training data.

[0053] It can be seen that in this example, the server optimizes the machine learning process by detecting the existence of available historical learning tasks and available historical training sets in the database, and then merging the historical learning tasks and historical training sets, thereby avoiding reducing the efficiency of machine learning due to repeated creation of training sets, and ensuring the effect of machine learning by setting the minimum number of training data in the first training set, avoiding deviations in learning effects due to insufficient training samples resulting from a small number of training data.

[0054] In a possible example, after merging the original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks through the task merging instruction to obtain the first training set, the method further includes: receiving a task integration instruction from a client; reselecting T historical learning tasks and merging the T historical learning tasks through the task integration instruction to obtain the target learning task; merging the original training set with X training data and the historical training set with S historical training data corresponding to the T historical learning tasks through the task integration instruction to obtain the first training set, wherein the value of X+S is greater than the target preset value.

[0055] Among them, users can query the merged and unmerged historical learning tasks through the client, and can reselect historical learning tasks to merge. Figure 5 , Figure 5 This is a schematic diagram of the third client interaction page provided in the embodiment of the present application, such as Figure 5 As shown, the user can reselect the T historical learning tasks to be merged and check the checkbox in front, and then click the "Confirm" control to make the client send the task integration instruction to the server. The server merges the reselected T historical learning tasks according to the task integration instruction to obtain the target learning task, and merges the original training set with X training data and the historical training set with S historical training data corresponding to the T historical learning tasks to obtain the first training set with (X+S) training data.

[0056] It can be seen that in this example, the user can reselect the historical learning tasks to be merged, so that the server can reintegrate the training data in the training set, ensuring that the machine learning process is carried out in accordance with the user's expectations, thereby increasing the flexibility of the solution.

[0057] In a possible example, determining a target learning task and a first training set corresponding to the current operating condition information based on the current operating condition information and reference operating condition information pre-stored in a learning training set database includes: detecting that there is no historical learning task corresponding to the current operating condition information in the reference operating condition information, but there are Z historical training data corresponding to the current operating condition information; obtaining an original training set with X training data; receiving a task creation instruction and a training data use confirmation instruction from a client; creating the target learning task according to the task creation instruction; and merging the original training set with X training data and the Z historical training data according to the training data use confirmation instruction to obtain the first training set, wherein the value of X+Z is greater than the target preset value.

[0058] Among them, the server detects that there are no historical learning tasks in the learning training set database, but there are historical training data corresponding to the current working condition information. At this time, in order to improve the efficiency of machine learning, these historical training data will be fully utilized for training. Since there are no historical learning tasks that can be merged, it is still necessary to create a new learning task to associate the training data and working conditions. In addition, since the historical training data may not meet the quantity requirements, or the department’s historical training data is unavailable, the original training set will still be obtained when a new learning task is created. At this time, the X training data in the original training set can be adjusted according to the size of the number of historical training data Z, as long as the number of training data in the first training set (X+Z) is greater than the target preset value. For example, please refer to Figure 6 , Figure 6 This is a schematic diagram of the fourth client interaction page provided in the embodiment of the present application. The server sends a message to the client, and a pop-up window such as Figure 6 The interactive interface shown in the figure displays a "Training Data Usage Confirmation" status bar: "Z batches of training data are detected to be available, are you sure to use these training data?" The user clicks the "Add" control to enable the client to send the training data usage confirmation instruction to the server. The server merges the original training set with X training data with the Z historical training data according to the instruction to obtain the first training set with (X+Z) training data. Then, the "Create Task" control is clicked to enable the client to send the task creation instruction to the server. The server creates the target learning task according to the instruction and associates it with the current working condition information.

[0059] It can be seen that in this example, the server detects that there is no available historical learning task in the database but there is available historical training data, merges the original training set and the historical training data to optimize the machine learning process, avoids reducing the efficiency of machine learning due to repeated creation of training sets, and ensures the effect of machine learning by setting the minimum number of training data in the first training set, avoiding deviations in learning effects due to insufficient training samples caused by a small number of training data.

[0060] In one possible example, the method of clearing the problematic first training data set from the first training set to obtain the second training set includes: setting the first training set to an available state, the first training set including M training data, M being a positive integer and greater than the target preset value, the M training data including M process parameters and production results corresponding to the M process parameters; obtaining a target production batch corresponding to the current operating condition information; determining that the production results corresponding to the M process parameters include A abnormal results, the abnormal results being used to indicate that there is at least one parameter in the production results that has a value that is not within a preset target value range; and clearing the A abnormal results and the process parameters corresponding to the A abnormal results in the first training set to obtain the second training set.

[0061] The abnormal result refers to the fact that one or more values ​​of the parameters included in a production result do not conform to the target value interval corresponding to the value, and the production result is an abnormal production result. The first training set is set to an available state to ensure that it can be smoothly input into the prediction model. For example, see Figure 7 , Figure 7 is a schematic diagram of clearing abnormal training data provided by an embodiment of the present application, such as Figure 7 As shown, the target production batch is a batch of the production process of product A, which is obtained by processing the component B as the initial material and processing the device C. The production batch number is recorded as 000111, and the operating condition is the parameter setting standard of the device C. The operating condition number is recorded as 121112111. Under this operating condition, multiple trainings are performed with different process parameters. For example, 9 trainings are performed. Then, there are 9 training data in the first training set. The 9 training data are marked and numbered by the server to obtain nine numbers N1, N2, ..., N9. It is determined whether the value of the parameter K included in the production result corresponding to each number is within the preset target value range. If the value of the parameter K of N2, N5, and N7 is not within the preset quality qualified range, the training data represented by N2, N5, and N7 are abnormal data. The training data represented by N2, N5, and N7 are removed, and the data set consisting of the remaining 6 training data is the second training set. It should be noted that if there is a parameter L corresponding to a production result that is not within the preset target value range, then the production result is also an abnormal result.

[0062] In a specific implementation, it is determined that the production results corresponding to the M process parameters include A abnormal results, and the A abnormal results are all those whose values ​​of parameter K are not within the preset target value interval. The method also includes: determining the target production equipment corresponding to the target parameter whose parameter value is not within the preset target interval; comparing the process parameters of the target production equipment corresponding to MA normal results with the process parameters of the target production equipment corresponding to the A abnormal results to determine the target process parameter with the abnormality; obtaining the first process parameters corresponding to other production equipment in the MA normal results and the first process parameters corresponding to other production equipment corresponding to the A abnormal results whose parameter values ​​are less than the preset value, and the first process parameters corresponding to N normal results The parameter does not include the target process parameter; obtain the parameter values ​​of the target process parameter corresponding to the N normal results to obtain the parameter range of the target process parameter, and the parameter values ​​of the target process parameter corresponding to the A abnormal results are not within the parameter range; obtain the target production batch corresponding to the interval endpoint value of the parameter range; determine the target production result corresponding to the target production batch; determine the influence trend of the target process parameter on the parameter value of the target product; determine a reference first process parameter based on the first process parameter; generate a reference result based on the target production result, the reference target process parameter, the reference first process parameter and the influence trend, and the value of the reference target process parameter is within the parameter range.

[0063] For example, in the above 9 training data, the production result numbered N2 is an abnormal result, and the value of parameter K of N2 is not within the preset target value range, and parameter K is the parameter corresponding to equipment P, and the process parameter corresponding to equipment P is Z. At this time, the training data with the smallest difference with the first process parameter of N2 among the other 8 training data is determined, and the first process parameter includes all parameters except the target process parameter corresponding to equipment P among the process parameters used for production. If all the parameters used for production include two, the process parameter corresponding to equipment P is parameter Z, and the process parameters corresponding to other equipment are parameter X, and the value of parameter X in the production batch corresponding to N2 is 40, then in the 9 training data, the value of parameter X is between 39.8-40.2, and the production batches between them are N3 and N4, then the parameter range can be determined according to the value of parameter K corresponding to the production of N3 and N4. For example, the value of parameter Z in N3 is 83, and the value of parameter Z in N4 is 78. At this time, if the value of parameter Z corresponding to batch N2 is 80, then the parameter range of parameter Z can be 80-83 or 78-80. Therefore, the reference first process parameter corresponding to the parameter X can be determined in the interval of 39.8-40.2, and the reference target process parameter can be determined in the interval of 80-83 or 78-80. Then, according to the reference first process parameters determined in these three intervals, multiple data groups are formed by combination, and the influence trend is determined. The influence trend can be that the larger the value of the process parameter Z, the larger the value corresponding to the production result. Then, according to the corresponding first reference process parameter and influence trend in each data group, and the production results corresponding to the two batches N3 and N4, the corresponding reference production results are generated to obtain reference training data, and the reference training data is added to the second training set.

[0064] It can be seen that in this example, the accuracy of machine learning is improved by removing abnormal data in the first training set in advance.

[0065] In one possible example, after storing the target learning result in the learning training set database to obtain an updated learning training set database, the method further includes: obtaining a query request from a client; determining whether a learning task exists in the current production batch based on the query request; querying the learning training set database to obtain a training data set corresponding to the learning task; and sending the training data set to the client, wherein the training data set is used to configure parameter values ​​of production equipment used to perform the current production batch.

[0066] Among them, after the machine learning process corresponding to the working condition is completed, the learning results are input into the learning training set database to update the data in the learning training set database for subsequent repeated use in machine learning. Specifically, in a certain production process, it is necessary to configure the parameters of the production equipment of the current production batch in the production process. At this time, it is preferred to query from the database whether there is a learning task for the current production batch. If there is a learning task, the training data set corresponding to the learning task is obtained first, and the parameter configuration of the production equipment is completed according to the training data set. In this way, the original purpose of machine learning can be achieved, that is, multiple training data that meet the standards are obtained through machine learning. These training data are the parameter values ​​of various production parameters in the actual production process, so that the multiple training data that meet the standards can be directly applied in the actual production process to improve production efficiency.

[0067] It can be seen that in this example, in the actual production process, by first detecting whether there are available learning tasks in the current production batch, the training data in the learning task is directly pulled for parameter configuration, the parameter configuration process is optimized, and production efficiency is improved.

[0068] In accordance with the above-mentioned embodiments, please refer to Figure 8a , Figure 8a is a functional unit block diagram of a data control device provided in an embodiment of the present application, wherein the device is applied to a server, such as Figure 8a As shown, the data control device 80 includes: a determination unit 801, which is used to determine the target learning task and the first training set corresponding to the current operating condition information according to the current operating condition information and the reference operating condition information pre-stored in the learning training set database, the target learning task is used to indicate the learning of the production parameters under the current operating condition through a preset model, and the first training set is used to indicate the training data set used for learning in the target learning task; a clearing unit 802, which is used to clear the first training data set with problems from the first training set to obtain a second training set; a training unit 803, which is used to train according to the second training set and the target learning task to obtain a target learning result; a storage unit 804, which is used to store the target learning result in the learning training set database to obtain an updated learning training set database.

[0069] In a possible example, in terms of determining the target learning task and the first training set corresponding to the current operating condition information based on the current operating condition information and the reference operating condition information pre-stored in the learning training set database, the determination unit 801 is specifically used to: detect that there is no historical learning task corresponding to the current operating condition information in the reference operating condition information, and there is no historical training data corresponding to the current operating condition information; obtain an original training set with X training data; receive a task creation instruction from a client; create the target learning task according to the task creation instruction; determine the first training set according to the task creation instruction and the original training set with X training data, where the value of X is greater than a target preset value.

[0070] In a possible example, in terms of determining the target learning task and the first training set corresponding to the current operating condition information based on the current operating condition information and the reference operating condition information pre-stored in the learning training set database, the determination unit 801 is specifically used to: detect that there are multiple historical learning tasks corresponding to the current operating condition information and historical training sets corresponding to the historical learning tasks in the reference operating condition information; obtain an original training set with X training data; receive a task merging instruction and a task selection instruction from the client; select N historical learning tasks through the task selection instruction; merge the N historical learning tasks through the task merging instruction to obtain the target learning task; merge the original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks through the task merging instruction to obtain the first training set, wherein the value of X+Y is greater than the target preset value.

[0071] In a possible example, after the original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks are merged through the task merging instruction to obtain the first training set, the determination unit 801 is further used to: receive a task integration instruction from the client; reselect T historical learning tasks and merge the T historical learning tasks through the task integration instruction to obtain the target learning task; merge the original training set with X training data and the historical training set with S historical training data corresponding to the T historical learning tasks through the task integration instruction to obtain the first training set, wherein the value of X+S is greater than the target preset value.

[0072] In a possible example, in terms of determining the target learning task and the first training set corresponding to the current operating condition information based on the current operating condition information and the reference operating condition information pre-stored in the learning training set database, the determination unit 801 is specifically used to: detect that there is no historical learning task corresponding to the current operating condition information in the reference operating condition information, but there are Z historical training data corresponding to the current operating condition information; obtain an original training set with X training data; receive a task creation instruction and a training data use confirmation instruction from the client; create the target learning task according to the task creation instruction; merge the original training set with X training data and the Z historical training data according to the training data use confirmation instruction to obtain the first training set, where the value of X+Z is greater than the target preset value.

[0073] In one possible example, in terms of clearing the problematic first training data set from the first training set to obtain the second training set, the clearing unit 802 is specifically used to: set the first training set to an available state, the first training set includes M training data, M is a positive integer and is greater than the target preset value, and the M training data include M process parameters and production results corresponding to the M process parameters; obtain the target production batch corresponding to the current operating condition information; determine that the production results corresponding to the M process parameters include A abnormal results, and the abnormal results are used to indicate that there is at least one parameter in the production results that has a value that is not within the preset target value range; clear the A abnormal results in the first training set and the process parameters corresponding to the A abnormal results to obtain the second training set.

[0074] In one possible example, after storing the target learning result in the learning training set database to obtain an updated learning training set database, the data management device 80 is also used to: obtain a query request from a client; determine whether a learning task exists in the current production batch based on the query request; query the learning training set database to obtain a training data set corresponding to the learning task; and send the training data set to the client, where the training data set is used to configure parameter values ​​of production equipment used to perform the current production batch.

[0075] It can be understood that since the method embodiment and the device embodiment are different presentation forms of the same technical concept, the content of the method embodiment part in this application should be synchronously adapted to the device embodiment part and will not be repeated here.

[0076] In the case of an integrated unit, such as Figure 8b As shown, Figure 8b This is a block diagram of the functional units of another data control device provided in an embodiment of the present application. Figure 8bIn the embodiment, the data control device 81 includes: a processing module 812 and a communication module 811. The processing module 812 is used to control and manage the actions of the data control device, for example, to execute the steps of the determination unit 801, the clearing unit 802, the training unit 803 and the storage unit 804, and / or to execute other processes of the technology described herein. The communication module 811 is used to support the interaction between the data control device and other devices. Figure 8b As shown, the data control device may further include a storage module 813, and the storage module 813 is used to store program codes and data of the data control device. The data control device 81 may be the data control device 80 described above.

[0077] Among them, the processing module 812 can be a processor or a controller, for example, a central processing unit (CPU), a general processor, a digital signal processor (DSP), an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It can implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements a computing function, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like. The communication module 811 can be a transceiver, an RF circuit or a communication interface, and the like. The storage module 813 can be a memory.

[0078] Among them, all relevant contents of each scenario involved in the above method embodiment can be referred to the functional description of the corresponding functional module, and will not be repeated here. The above data control device 81 can execute the above Figure 2 The data governance approach shown.

[0079] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0080] Fig. 9 is a structural block diagram of a server provided in an embodiment of the present application. Fig. 9 As shown, the server 900 may include one or more of the following components: a processor 901, a memory 902 coupled to the processor 901, wherein the memory 902 may store one or more computer programs, and the one or more computer programs may be configured to implement the methods described in the above embodiments when executed by one or more processors 901. The server 900 may be the aforementioned server 11.

[0081] The processor 901 may include one or more processing cores. The processor 901 uses various interfaces and lines to connect the various parts of the entire server 900, and executes various functions and processes data of the server 900 by running or executing instructions, programs, code sets or instruction sets stored in the memory 902, and calling data stored in the memory 902. Optionally, the processor 901 can be implemented in at least one hardware form of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 901 can integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing display content; and the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 901, but may be implemented separately through a communication chip.

[0082] The memory 902 may include a random access memory (RAM) or a read-only memory (ROM). The memory 902 may be used to store instructions, programs, codes, code sets or instruction sets. The memory 902 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc. The data storage area may also store data created by the server 900 during use, etc.

[0083] It is understandable that the server 900 may include more or fewer structural elements than those in the above structural block diagram, for example, including a power module, physical buttons, a WiFi (Wireless Fidelity) module, a speaker, a Bluetooth module, a sensor, etc., which are not limited here.

[0084] An embodiment of the present application also provides a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute part or all of the steps of any method recorded in the above method embodiments.

[0085] The present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute some or all of the steps of any method described in the above method embodiment. The computer program product may be a software installation package.

[0086] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0087] In the several embodiments provided in the present application, it should be understood that the disclosed methods, devices and systems can be implemented in other ways. For example, the device embodiments described above are merely schematic; for example, the division of the units is only a logical function division, and there may be other division methods in actual implementation; for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0088] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0089] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may be physically included separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0090] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform some steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a volatile memory or a non-volatile memory. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM) and direct RAM bus random access memory (DR RAM), among other media that can store program code.

[0091] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions without departing from the spirit and scope of the present invention, and can make various changes and modifications, including the combination of the above-mentioned different functions and implementation steps, including software and hardware implementation methods, all of which are within the scope of protection of the present invention.

Claims

1. A data control method, characterized in that: The method comprises: Determine a target learning task and a first training set corresponding to the current working condition information according to the current working condition information and the reference working condition information pre-stored in the learning training set database, wherein the target learning task is used to indicate learning the production parameters under the current working condition through a preset model, and the first training set is used to indicate a training data set used for learning in the target learning task; the learning training set database stores training records of historical working conditions, and each historical working condition has a unique working condition number for distinguishing different working conditions; Remove the problematic first training data set from the first training set to obtain a second training set; Performing training according to the second training set and the target learning task to obtain a target learning result; Storing the target learning result in the learning training set database to obtain an updated learning training set database; The step of determining the target learning task and the first training set corresponding to the current working condition information according to the current working condition information and the reference working condition information pre-stored in the learning training set database includes: It is detected that there are a plurality of historical learning tasks corresponding to the current operating condition information and a plurality of historical training sets corresponding to the historical learning tasks in the reference operating condition information; Get the original training set with X training data; Receive task merging instructions and task selection instructions from the client; Selecting N historical learning tasks through the task selection instruction; Merging the N historical learning tasks through the task merging instruction to obtain the target learning task; The original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks are merged through the task merging instruction to obtain the first training set, wherein the value of X+Y is greater than the target preset value.

2. The method according to claim 1, characterized in that: The step of determining a target learning task and a first training set corresponding to the current operating condition information according to the current operating condition information and reference operating condition information pre-stored in a learning training set database includes: It is detected that there is no historical learning task corresponding to the current operating condition information in the reference operating condition information, and there is no historical training data corresponding to the current operating condition information; Get the original training set with X training data; Receive task creation instructions from the client; Creating the target learning task according to the task creation instruction; The first training set is determined according to the task creation instruction and the original training set with X training data, wherein the value of X is greater than a target preset value.

3. The method according to claim 1, characterized in that: After merging the original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks through the task merging instruction to obtain the first training set, the method further includes: Receive task integration instructions from the client; Reselecting T historical learning tasks and merging the T historical learning tasks through the task integration instruction to obtain the target learning task; The task integration instruction is used to merge the original training set with the X training data and the historical training set with the S historical training data corresponding to the T historical learning tasks to obtain the first training set, wherein the value of X+S is greater than the target preset value.

4. The method according to claim 1, characterized in that: The step of determining a target learning task and a first training set corresponding to the current operating condition information according to the current operating condition information and reference operating condition information pre-stored in a learning training set database includes: It is detected that there is no historical learning task corresponding to the current operating condition information in the reference operating condition information, but there are Z historical training data corresponding to the current operating condition information; Get the original training set with X training data; Receive task creation instructions and training data usage confirmation instructions from the client; Creating the target learning task according to the task creation instruction; The original training set with the X training data and the Z historical training data are merged according to the training data use confirmation instruction to obtain the first training set, wherein a value of X+Z is greater than a target preset value.

5. The method according to any one of claims 1 to 4, characterized in that: The step of removing the problematic first training data set from the first training set to obtain a second training set includes: Setting the first training set to an available state, the first training set includes M training data, where M is a positive integer and is greater than the target preset value, and the M training data include M process parameters and production results corresponding to the M process parameters; Determining that the production results corresponding to the M process parameters include A abnormal results, where the abnormal results are used to indicate that a value of at least one parameter in the production results is not within a preset target value range; A abnormal results and process parameters corresponding to the A abnormal results in the first training set are removed to obtain the second training set.

6. The method according to any one of claims 1 to 4, characterized in that: After storing the target learning result in the learning training set database to obtain an updated learning training set database, the method further includes: Get the query request from the client; Determining, according to the query request, that a learning task exists in the current production batch; Querying the learning training set database to obtain a training data set corresponding to the learning task; The training data set is sent to the client, where the training data set is used to configure parameter values ​​of a production device used to perform the current production batch.

7. The method according to claim 5, characterized in that: After storing the target learning result in the learning training set database to obtain an updated learning training set database, the method further includes: Get the query request from the client; Determining, according to the query request, that a learning task exists in the current production batch; Querying the learning training set database to obtain a training data set corresponding to the learning task; The training data set is sent to the client, where the training data set is used to configure parameter values ​​of a production device used to perform the current production batch.

8. A data control device, characterized in that: The device comprises: A determination unit, used to determine a target learning task and a first training set corresponding to the current working condition information according to the current working condition information and the reference working condition information pre-stored in the learning training set database, wherein the target learning task is used to indicate learning of production parameters under the current working condition through a preset model, and the first training set is used to indicate a training data set used for learning in the target learning task; the learning training set database stores training records of historical working conditions, and each historical working condition has a unique working condition number for distinguishing different working conditions; a clearing unit, configured to clear the first training data set with problems from the first training set to obtain a second training set; A training unit, configured to perform training according to the second training set and the target learning task to obtain a target learning result; A storage unit, used for storing the target learning result in the learning training set database to obtain an updated learning training set database; The step of determining the target learning task and the first training set corresponding to the current working condition information according to the current working condition information and the reference working condition information pre-stored in the learning training set database includes: It is detected that there are a plurality of historical learning tasks corresponding to the current operating condition information and a plurality of historical training sets corresponding to the historical learning tasks in the reference operating condition information; Get the original training set with X training data; Receive task merging instructions and task selection instructions from the client; Selecting N historical learning tasks through the task selection instruction; Merging the N historical learning tasks through the task merging instruction to obtain the target learning task; The original training set with X training data and the historical training set with Y historical training data corresponding to the N historical learning tasks are merged through the task merging instruction to obtain the first training set, wherein the value of X+Y is greater than the target preset value.

9. A server, characterized in that: The method comprises a processor, a memory and a communication interface, wherein one or more programs are stored in the memory and configured to be executed by the processor, wherein the programs include instructions for executing the steps in the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program enables a computer to execute the method according to any one of claims 1 to 7.

11. A computer program product, characterized in that The computer program product enables a computer to execute the method according to any one of claims 1 to 7.

Images