CN114936807A

CN114936807A - Machine learning progress control method, and related apparatus, storage medium, and program product

Info

Publication number: CN114936807A
Application number: CN202210787265.5A
Authority: CN
Inventors: 郭传亮
Original assignee: Hope Zhizhou Technology Shenzhen Co ltd
Current assignee: Hope Zhizhou Technology Shenzhen Co ltd
Priority date: 2022-03-11
Filing date: 2022-03-11
Publication date: 2022-08-23
Also published as: CN114418456A; CN114418456B

Abstract

The embodiment of the application provides a machine learning progress control method based on working conditions, a related device, a storage medium and a program product, wherein the method comprises the following steps: acquiring a progress query request from the electronic equipment; determining the current learning condition of the target working condition according to the target working condition code; determining whether a preliminary learning stage exists in the target working condition according to the current learning condition; if the target working condition exists in the pre-learning stage, sending the current learning condition and prompt information to the electronic equipment, wherein the prompt information is used for prompting a user to create a task of the target working condition in the pre-learning stage; and if not, sending the current learning condition to the electronic equipment. Therefore, the training stage is automatically identified through the working condition, and the user is assisted to complete the whole process of machine learning by the information display and prompt message sending mode, so that the machine learning efficiency is greatly improved, the user can quickly know the current learning progress, and the decision difficulty of the user is reduced.

Description

Machine learning progress control method, and related apparatus, storage medium, and program product

Technical Field

The application belongs to the field of general data processing of the Internet industry, and particularly relates to a machine learning progress control method based on working conditions, a related device, a storage medium and a program product.

Background

At present, the trend of more and more popular is to help to optimize process parameters through machine learning training production, but in the traditional machine learning training, learning process parameter data are generated through the plan arrangement of personnel and then arranged in a production flow, the progress and the quality of the process of trial production training are managed manually, meanwhile, the training has a plurality of stages such as training and learning, the verification needs personnel to make decisions, the management and control are coordinated, on one hand, the efficiency is low, the training depends on the skill experience of the personnel, and meanwhile, errors are easy to occur.

Disclosure of Invention

The embodiment of the application provides a machine learning progress control method based on working conditions, a related device, a storage medium and a program product, so that the machine learning efficiency is improved, and the convenience of controlling the whole machine learning process by a user is improved.

In a first aspect, an embodiment of the present application provides a working condition-based machine learning progress control method, which is applied to a server of a progress control system, where the progress control system includes the server and an electronic device, and the method includes:

acquiring a progress query request from the electronic equipment, wherein the progress query request comprises a target working condition code, and the target working condition code is used for indicating a target working condition;

determining the current learning condition of the target working condition according to the target working condition code, wherein the current learning condition is used for indicating the current learning condition of the target working condition in each learning stage;

determining whether a preliminary learning stage exists in the target working condition according to the current learning condition, wherein the preliminary learning stage is a learning stage which can be entered but is not entered at present in the target working condition;

if the pre-learning stage exists, sending the current learning condition and prompt information to the electronic equipment, wherein the prompt information is used for prompting a user to create a task of the target working condition in the pre-learning stage;

and if the preliminary learning stage does not exist, sending the current learning condition to the electronic equipment.

In a second aspect, an embodiment of the present application provides a working condition-based machine learning progress control method, which is applied to an electronic device of a progress control system, where the progress control system includes a server and the electronic device, and the method includes:

sending a progress query request to the server, wherein the progress query request comprises a target working condition code, the target working condition code is used for indicating a target working condition, the target working condition is used for indicating a current learning condition, the current learning condition is used for indicating the current learning condition of the target working condition in each learning stage, the current learning condition is used for determining whether a pre-learning stage exists in the target working condition, and the pre-learning stage is a learning stage which can be entered but is not entered in the target working condition at present;

if the pre-learning stage exists, acquiring the current learning condition and prompt information sent by the server, wherein the prompt information is used for prompting a user to create a task of the target working condition in the pre-learning stage;

and if the preliminary learning stage does not exist, acquiring the current learning situation from the server.

In a third aspect, an embodiment of the present application provides a working condition-based machine learning progress control apparatus, which is applied to a server of a progress control system, where the progress control system includes the server and an electronic device, and the apparatus includes:

the device comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring a progress query request from the electronic equipment, the progress query request comprises a target working condition code, and the target working condition code is used for indicating a target working condition;

the first determining unit is used for determining the current learning condition of the target working condition according to the target working condition code, wherein the current learning condition is used for indicating the current learning condition of the target working condition in each learning stage;

a second determining unit, configured to determine whether a preliminary learning stage exists in the target operating condition according to the current learning condition, where the preliminary learning stage is a learning stage that the target operating condition may currently enter but has not yet entered;

a first sending unit, configured to send the current learning condition and prompt information to the electronic device if the pre-learning stage exists, where the prompt information is used to prompt a user to create a task of the target working condition in the pre-learning stage;

a second sending unit, configured to send the current learning situation to the electronic device if the preliminary learning stage does not exist.

In a fourth aspect, an embodiment of the present application provides a machine learning progress control apparatus based on operating conditions, which is applied to an electronic device of a progress control system, where the progress control system includes a server and the electronic device, the apparatus includes:

a sending unit, configured to send a progress query request to the server, where the progress query request includes a target working condition code, the target working condition code is used to indicate a target working condition, the target working condition is used to indicate a current learning condition, the current learning condition is used to indicate a learning condition of the target working condition in each learning stage, the current learning condition is used to determine whether a preliminary learning stage exists in the target working condition, and the preliminary learning stage is a learning stage that the target working condition can currently enter but has not yet entered;

a first obtaining unit, configured to obtain, if the pre-learning stage exists, the current learning condition and prompt information sent by the server, where the prompt information is used to prompt a user to create a task of the target working condition in the pre-learning stage;

a second obtaining unit, configured to obtain the current learning situation from the server if the preliminary learning stage does not exist.

In a fifth aspect, embodiments of the present application provide a server comprising a processor, a memory, a communication interface, and one or more programs stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps of the first aspect of embodiments of the present application.

In a sixth aspect, embodiments of the present application provide an electronic device, comprising a processor, a memory, a communication interface, and one or more programs, stored in the memory and configured to be executed by the processor, the programs including instructions for performing the steps of the second aspect of embodiments of the present application.

In a seventh aspect, this application embodiment provides a computer program product, where the computer program product includes a non-transitory computer-readable storage medium storing a computer program, where the computer program is operable to cause a computer to perform some or all of the steps as described in the first aspect or the second aspect of this application embodiment. The computer program product may be a software installation package.

It can be seen that, in the embodiment of the present application, a server first obtains a progress query request from an electronic device, then determines a current learning condition of the target working condition according to the target working condition code, then determines whether a preliminary learning stage exists in the target working condition according to the current learning condition, and if the preliminary learning stage exists, sends the current learning condition and prompt information to the electronic device, where the prompt information is used to prompt a user to create a task of the target working condition in the preliminary learning stage, and if the preliminary learning stage does not exist, sends the current learning condition to the electronic device. Therefore, the training stage is automatically identified through the working condition, and the user is assisted to complete the whole process of machine learning by the information display and prompt message sending mode, so that the machine learning efficiency is greatly improved, the user can quickly know the current learning progress, the decision-making difficulty of the user is reduced, and the use experience of the user is improved.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a schematic structural diagram of a progress management and control system according to an embodiment of the present disclosure;

fig. 2 is a schematic structural diagram of a server provided in an embodiment of the present application;

fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application;

fig. 4 is a schematic flowchart of a method for managing and controlling a machine learning progress based on a working condition according to an embodiment of the present application;

FIG. 5 is a schematic view of a machine learning progress control interface according to an embodiment of the present disclosure;

FIG. 6 is a schematic diagram of another machine learning progress control interface provided in the embodiment of the present application;

FIG. 7 is a schematic diagram of another machine learning progress control interface provided in the embodiment of the present application;

fig. 8 is a block diagram of functional units of a machine learning progress control apparatus based on a working condition according to an embodiment of the present application;

fig. 9 is a block diagram of functional units of another working condition-based machine learning progress control apparatus according to an embodiment of the present disclosure;

fig. 10 is a block diagram illustrating functional units of another working condition-based machine learning progress control apparatus according to an embodiment of the present disclosure;

fig. 11 is a block diagram illustrating functional units of another condition-based machine learning progress control apparatus according to an embodiment of the present disclosure.

Detailed Description

In order to make the technical solutions of the present application better understood, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

The terms "first," "second," and the like in the description and claims of the present application and in the above-described drawings are used for distinguishing between different objects and not for describing a particular order. Furthermore, the terms "include" and "have," as well as any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements but may alternatively include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.

In process manufacturing industries such as: in the chemical industry, metallurgy, steel and other industries, the trial production of a plurality of new products and the new working condition of old products need to optimize the climbing process through the production process, along with the development of artificial intelligence technology, the current production through machine learning training helps to optimize the process parameters, but the traditional machine learning training is to generate learning process parameter data through the plan arrangement of personnel, arrange the learning process parameter data into the production process, manually manage the progress and quality of the process of trial production training, and simultaneously train a plurality of stages, such as training, learning, verification and personnel decision making, coordinate management and control, on one hand, the efficiency is low, the skill experience of the personnel is also relied on, and meanwhile, errors are easy to occur.

In view of the above problems, embodiments of the present application provide a machine learning progress control method and a related apparatus based on operating conditions, and the following describes embodiments of the present application in detail with reference to the accompanying drawings.

Referring to fig. 1, fig. 1 is a schematic structural diagram of a progress control system according to an embodiment of the present disclosure. As shown in the figure, the progress management and control system 10 includes a server 101 and an electronic device 102, where the electronic device 102 is configured to send a progress query request to the server, and the server 101 obtains progress information according to the query request of the electronic device 102 and feeds the progress back to the electronic device 102 for a user to view. And the user can also send change information to the server 101 through the electronic device 102 to update the current progress or change information in the machine learning process.

The server 101 structure is shown in fig. 2, where the server 101 includes a processor 120, a memory 130, a communication interface 140, and one or more programs 131, where the one or more programs 131 are stored in the memory 130 and configured to be executed by the processor 120, and the one or more programs 131 include instructions for performing any of the steps of the method embodiments described below. In a specific implementation, the processor 120 is configured to perform any one of the steps performed by the server in the method embodiments described below, and when performing a data transmission operation such as receiving, optionally invokes the communication interface 140 to complete the corresponding operation. It should be noted that, in the present solution, the storage manner of the servers may be distributed, and multiple servers in different production lines, workshops and quality laboratories are used, and then the database storage technology of the block chain is used to record the production information of each batch of materials in real time, and the production information is linked and stored in the order of the production time of different production equipment.

The structure of the electronic device 102 is shown in fig. 3, where the electronic device 102 includes a processor 220, a memory 230, a communication interface 240, and one or more programs 231, where the one or more programs 231 are stored in the memory 230 and configured to be executed by the processor 220, and the one or more programs 231 include instructions for performing any of the steps of the method embodiments described below. In a specific implementation, the processor 220 is configured to perform any one of the steps performed by the electronic device in the method embodiments described below, and when performing a data transmission operation such as receiving, optionally invokes the communication interface 240 to complete the corresponding operation.

Referring to fig. 4, fig. 4 is a schematic flowchart illustrating a method for managing and controlling a machine learning progress based on a working condition according to an embodiment of the present disclosure. As shown in the figure, the working condition-based machine learning progress control method is applied to a server of a progress control system, the progress control system comprises the server and an electronic device, and the method comprises the following steps:

s201, obtaining a progress query request from the electronic equipment, wherein the progress query request comprises a target working condition code, and the target working condition code is used for indicating a target working condition.

Each working condition code corresponds to one type of working condition, and the working condition is used for indicating the current production condition, including the production environment condition, the production equipment condition, the production personnel condition and the like. And establishing a machine learning task aiming at each working condition code so as to control the production parameters under the working conditions according to the result obtained by machine learning.

S202, determining the current learning condition of the target working condition according to the target working condition code, wherein the current learning condition is used for indicating the current learning condition of the target working condition in each learning stage.

The process of the whole machine learning is multiple and consumes long time, and meanwhile, the working conditions corresponding to a plurality of working condition codes can be used for machine learning, so that the user can not quickly perform effective and accurate management and control on each machine learning process. The current learning situation includes the case of the learning phase in which learning has been completed, the learning phase in which learning is in progress, and the learning phase in which learning has not yet been performed. A user can know which steps are performed and which steps need to be performed in the machine learning corresponding to the current working condition code at a glance through the current learning condition sent by the server, so that the user can conveniently and quickly know the current learning progress and arrange subsequent production work according to the current learning progress.

S203, determining whether a preliminary learning stage exists in the target working condition according to the current learning condition, wherein the preliminary learning stage is a learning stage which can be entered but is not entered currently in the target working condition.

Wherein, since the machine learning comprises a plurality of learning stages in total and each learning stage is closely related, the following learning stage can be determined according to the learning condition. For a working condition, the working condition can correspond to a plurality of learning tasks, the plurality of learning tasks can comprise three states of being completed, being performed or newly built and not being performed, and only one learning task for the running state is included. A user may query, according to a current learning condition, learning information corresponding to a learned learning stage in a learning task that is currently learning, for example, the learned learning stage is a verification evaluation stage, and then the user may obtain a target value corresponding to each product parameter, data information such as a mean value, a standard deviation, and a Process Capability index (CPK) before learning corresponding to the product parameter, and data information such as a mean value, a standard deviation, and a CPK corresponding to a certain learning version corresponding to each product parameter.

S204, if the preparatory learning stage exists, sending the current learning condition and prompt information to the electronic equipment, wherein the prompt information is used for prompting a user to create a task of the target working condition in the preparatory learning stage;

and S205, if the preliminary learning stage does not exist, sending the current learning situation to the electronic equipment.

If the preliminary learning stage exists, the fact that the current production stage is not the last production stage of machine learning is meant, and at least one learning task is created when the current learning stage is completed or the initial stage of learning tasks is not performed under the working condition codes. For example, a plurality of learning tasks are associated with the condition code, where the learning task 5 is already completed in the current learning stage, so that a prompt message may be sent to the user to prompt the user to enter the next learning stage for the learning task. If there is no pre-learning stage, it means that the current learning stage is not completed, the next learning stage cannot be entered, or the information required by the next learning stage is not obtained, for example, at least 40 sets of training information are required in the learning stage, but the data result obtained according to the current training stage is not completed, the learning stage cannot be entered, and if the training data obtained in the training stage reaches the minimum value, even if the training stage is not completed, the learning stage can be entered.

It can be seen that, in this example, the server first obtains a progress query request from the electronic device, then determines a current learning condition of the target working condition according to the target working condition code, then determines whether the target working condition has a preliminary learning stage according to the current learning condition, and if so, sends the current learning condition and prompt information to the electronic device, where the prompt information is used to prompt a user to create a task of the target working condition in the preliminary learning stage, and if not, sends the current learning condition to the electronic device. Therefore, the training stage is automatically identified through the working condition, and the user is assisted to complete the whole process of machine learning by the information display and prompt message sending mode, so that the machine learning efficiency is greatly improved, the user can quickly know the current learning progress, the decision-making difficulty of the user is reduced, and the use experience of the user is improved.

In one possible example, the learning progress of the target working condition includes a learning task creating stage, a learning target confirming stage, a learning production scheduling stage, a training stage, a multi-target learning stage and a verification and evaluation stage, and the determining whether the target working condition has a pre-learning stage according to the current learning condition includes: determining a next learning stage of the current learning stage, wherein the learning task creating stage is an initial learning stage, the learning target confirming stage is the next learning stage of the learning task creating stage, the learning production scheduling stage is the next stage of the learning target confirming stage, the training stage is the next stage of the learning production scheduling stage, the multi-target learning stage is the next stage of the training stage, and the verification and evaluation stage is the next stage of the multi-target learning stage; determining whether the learning task of the target working condition in the current learning stage is finished; and if so, determining that the next learning stage of the current learning stage is a preparatory learning stage.

Wherein, as shown in fig. 5, the interface of "a product multi-objective optimization learning task navigation" in fig. 5 is shown, the machine learning may refer to multi-objective optimization learning, the multi-objective optimization learning totally includes 6 learning stages, respectively, in the stage of creating the learning task, a multi-objective machine learning task is created for a working condition corresponding to a certain working condition code, after the learning task is created, the learning objective of the learning is determined by a user in the stage of learning objective confirmation, then the learning production scheduling stage is entered according to the learning objective, a plurality of training data can be generated in this stage, then the training stage is entered after enough training data is obtained, a multi-objective optimization model is trained according to the training data, the training stage is entered after the training is completed, the multi-objective learning stage is entered through the multi-objective optimization model, and optimized product parameter values are obtained through calculation in this stage, and then, in the verification and evaluation stage, verification and evaluation are carried out on the basis of the product parameter values obtained by learning, the product parameter target values and the product parameter values before learning, and the versions which can be used for actual production are determined. The user can check the learning condition of the learning stage which is completed by learning according to the current learning condition, and when the preparatory learning stage exists, the user can be prompted to enter the next stage, for example, the user is prompted in a figure that the current working condition of the user does not have learning production scheduling, so that the user can click 'go scheduling' on the electronic equipment, and then the next learning stage can be entered. Particularly, the information of the current learning situation, for example, the information of each learning batch, including the progress data, the detection information, the operator, the inspector and the like, can be obtained by the block chain technology according to the time node of the block chain, so as to ensure the traceability and reliability of the information.

In the specific implementation, when the learning target is confirmed, it may be determined whether the target product has already been subjected to machine learning under other working conditions, if so, the learning target determined when the machine learning under other working conditions is obtained, and then the learning target is determined to be the learning target template of the machine learning, and the user only needs to make a corresponding change on the template, or the user can directly confirm that the learning target template is the learning target of the machine learning. The learning objective comprises the target value, the priority, the expected trend, the specification upper and lower limits and the like of each product parameter of the target product. When learning target confirmation is carried out, initial product parameters determined by the product in a laboratory stage can be obtained, and then the initial product parameters are directly imported into a learning target template.

Therefore, in the embodiment, the whole process of the multi-objective optimization learning is subjected to flow management and control and display, so that a user can determine the learning progress of machine learning at each time, follow-up operation can be performed according to prompts, convenience and learning efficiency of machine learning are improved, and user experience is improved.

In one possible example, the initial stage of the learning progress of the target operating condition is a stage of creating a learning task, the current learning condition of the target operating condition includes that the learning task is not created yet by the target operating condition, and after the current learning condition of the target operating condition is determined according to the target operating condition code, the method further includes: determining the working condition type of the target working condition according to the target working condition code; determining the number of recommended tasks according to the working condition types, wherein the number of recommended tasks is used for recommending the number of learning tasks which need to be created by the user in the learning task creating stage and aim at the target working condition; and sending the recommended task quantity to the electronic equipment.

The target working condition code may include contents of four aspects, that is, the production environment information, the production equipment information, the production material information, and the production personnel information may be determined according to the working condition code, and the current working condition type may be determined according to the contents of the four aspects. The current stage can be known through working condition coding and the progress information of machine learning, and the current stage belongs to the production type. For example, training production, multi-target machine learning production, mass production verification production and mass production can recommend the number of learning tasks required to be corresponding to the quality target achievement conditions of different production types. And the version state of the parameter database corresponding to the working condition codes can be determined to be in the mass production verification stage or the mass production stage of the trial production.

Therefore, in the embodiment, the corresponding learning task quantity is recommended according to different working condition types, so that the machine learning efficiency can be improved, the burden during machine learning is reduced, and the resource waste is reduced.

In one possible example, the determining the recommended number of tasks according to the operating condition type includes: determining the same-type working condition codes according to the target working condition codes, wherein the similarity between the same-type working condition codes and the target working condition codes is greater than a preset value; acquiring a preparation working condition, wherein the preparation working condition is a working condition that all learning stages are completed in the same working condition corresponding to the same working condition codes; determining the number of learning tasks corresponding to the preparation condition and the product comprehensive quality index corresponding to each learning task corresponding to the preparation condition; determining the quality distribution condition according to the product comprehensive quality index of each learning task; determining a recommended task number interval according to the quality distribution condition and the learning task number; and determining the recommended task quantity according to the recommended task quantity interval.

The operating condition codes can reflect the production conditions of multiple aspects, so that if the production conditions of only one aspect are different in the production conditions of the multiple aspects, only one digit in the same type of operating condition codes is different from the target operating condition codes. And the same type of working condition codes can comprise a plurality of working condition codes, for example, the target working condition codes and the same type of working condition codes are only different from each other in the condition of production personnel, the production personnel of the target working condition codes are A groups, the production personnel of the same type of working condition codes 1 are B groups, and the production personnel of the same type of working condition codes 2 are C groups. When a plurality of different types of similar working condition codes exist, for example, some similar working condition codes are different from the production environment condition of the target working condition code, and some similar working condition codes are different from the production personnel condition of the target working condition code, so that the obtained similar working condition codes can be grouped according to the conditions, for example, the obtained similar working condition codes totally include four production conditions, the production environment conditions are different into one group, the production equipment conditions are different into one group, the production material conditions are different into one group, the production personnel conditions are different into one group, and then the group with the largest number of working conditions which have completed all learning stages in the four groups of similar working condition codes is determined as the final similar working condition code.

In the specific implementation, if a factory has multiple production lines, M (M < N) production lines of N production line devices simultaneously produce and learn the working conditions of a similar product, and the similar working conditions encode (the production lines are different, but the device types are the same, and other working conditions are the same) learning tasks. When the working conditions are different only in production lines, namely (the similar working conditions, the equipment types and the like are the same), a plurality of similar working conditions exist, M (M < N) of N production lines learn simultaneously, if the learning standard reaching rate is high, the learning parameter consistency is good, parameters of M production lines can be marked and set to other production lines which are not learned, and the learning frequency is reduced.

When the preparation working condition comprises a plurality of, determining a recommended task quantity interval according to the quality distribution condition and the learning task quantity, including: determining first product comprehensive quality indexes corresponding to all learning tasks of each preparation working condition, then determining average product comprehensive quality indexes corresponding to the preparation working conditions according to the learning task quantity corresponding to each preparation working condition, then determining a total quality distribution interval according to the average product comprehensive quality indexes of all preparation working conditions, then equally dividing the total quality distribution interval, determining a recommended task quantity interval from the equally divided intervals, determining a target preparation working condition corresponding to the recommended task quantity interval according to the maximum preparation working condition quantity in the recommended task quantity interval, then determining the learning task quantity of each target preparation working condition, determining a prepared recommended task quantity according to the learning task quantity of each target preparation working condition, and determining the difference between the product comprehensive quality indexes corresponding to the recommended task quantity interval and the target value of the product comprehensive quality indexes, and determining the final recommended task quantity according to the gap and the prepared recommended task quantity. For example, if the product comprehensive quality index corresponding to the recommended task quantity interval includes a target value, the prepared recommended task quantity is determined to be the final recommended task quantity, if the recommended task quantity interval is smaller than the target value, the prepared recommended task quantity is determined to be smaller than the final recommended task quantity, and if the recommended task quantity interval is larger than the target value, the prepared recommended task quantity is determined to be larger than the final recommended task quantity. Or determining the recommendation proportion according to the gap value, and determining the final recommended task quantity according to the recommendation proportion and the prepared recommended task quantity.

Therefore, in the example, the recommended task number during the creation of the learning task is determined according to the working condition codes of the same type, so that the machine learning efficiency can be improved, the burden during the machine learning can be reduced, and the resource waste can be reduced.

In one possible example, the determining the current learning condition of the target condition according to the target condition code includes: determining a training total batch of the target working condition in the training stage; obtaining the production condition of each production flow aiming at the target working condition, wherein the production condition is used for indicating the production condition of the current production batch corresponding to the production flow; determining a completion progress corresponding to each production flow according to the training total batch and the current production batch; determining the training progress of the target working condition in the training stage according to the completion progress corresponding to each production flow; the sending the current learning situation to the electronic device includes: and sending the production condition and the training progress to the electronic equipment.

As shown in fig. 6, fig. 6 is a schematic view of another machine learning progress control interface provided in the embodiment of the present application. During training, the obtained training parameters can be used for generating production tasks of different batches, and when a user inquires the machine learning progress, the production conditions of different batches can be displayed. For example, the product of the first batch is currently produced in a certain process, and the completion progress of the process in the total training is what. Determining the completion progress may include: determining a total training batch, determining a current production batch of the target process, determining the number of produced finished batches according to the current production batch, determining the production progress of the current production batch, and determining the completion progress of the target process according to the production progress of the current production batch, the number of the finished production batches and the total training batch. For example, if the current production batch of the process 5 is the 21 st batch, it means that the process 5 has finished producing 20 batches, and if the production progress of the current production batch has finished 50%, the completion progress of the process 5 can be determined according to the obtained information.

Therefore, in the example, the production condition of each flow, the training progress and other information are displayed to the user in the training stage, so that the user can conveniently know the current machine learning progress, and the user management efficiency is improved.

In a possible example, the determining the training progress of the target operating condition in the training phase according to the completion progress corresponding to each production flow includes: determining a preparation production flow, wherein the completion progress corresponding to the preparation production flow is less than 1; numbering the preparation production flow according to a production sequence, wherein the larger the numbering sequence is, the later the production sequence is; determining a progress coefficient for each preliminary production flow according to the production sequence, wherein the more the production sequence of the preliminary production flow is, the higher the progress coefficient is; and determining the training progress of the target working condition in the training stage according to the progress coefficient and the completion progress corresponding to the preparation production flow.

Wherein, the completion progress equal to 1 means that the process has produced the content of all training batches. And the sum S of the progress coefficients of all the prepared production flows belongs to [0, N ], and N is the number of the production flows. The training progress of the current training stage can be visually displayed through the training progress bar.

In one possible example, when the current learning situation includes a multi-target learning phase, the determining the current learning situation of the target condition according to the target condition code includes: determining a learned turn of the target operating condition; acquiring product parameter information obtained by learning in each learning turn in the learned turns; sorting the learned turns according to the product parameter information; the sending the current learning situation to the electronic device includes: sending product parameter information corresponding to a target learning turn to the electronic equipment, wherein the target learning turn is a learned turn sequenced before a preset sequence; after the product parameter information corresponding to the target learning turn is sent to the electronic device, the method further includes: receiving a setting request from the electronic equipment for the versions of the target learning turns, wherein the learning parameter corresponding to the multi-target learning of each turn is one version; updating the version of the target learning round to a preset version according to the version setting request; and determining the preset version as a version which can enter a verification evaluation phase.

The method comprises the steps that multiple rounds of multi-target learning are needed for one working condition code, corresponding product parameter information is obtained after each round of multi-target learning, and the product parameter information comprises parameter information of multiple parameters. Ranking the learned turns according to the product parameter information includes: determining the priority of each parameter information in the plurality of parameter information, acquiring a parameter target interval corresponding to each parameter information, determining a comprehensive product index corresponding to each round of learning according to the parameter target interval and the priority, and determining an arrangement sequence according to the comprehensive product index. When the current learning situation is transmitted to the electronic device, only the learning results of the top-ranked learning round may be transmitted. And after receiving the request information from the electronic equipment, sending the learning results of other ordered learning rounds to the electronic equipment.

In specific implementation, version setting can be performed on the learning round with the front ranking, and after the learning information of the round is set as a preset version, verification evaluation can be performed according to the learning information of the round. With the increase of the learning rounds, the current learning round with the front ranking may be changed, that is, some versions of the learning round which may appear in the learning round with the front ranking are preset versions, and some versions of the learning round are not preset versions, so that the versions of the learning rounds can be set at this time. For example, in fig. 7, the version corresponding to the 5 th learning round may be added, and set as the preset version for verification evaluation.

In a specific implementation, as shown in fig. 7, learning information for each product parameter may also be obtained, including the parameter name, unit, target value, upper and lower specification limits, expected learning direction, and other contents of the product parameter, and the parameter value of the parameter learned each time is displayed in a line graph manner, so that the learning condition of each parameter can be conveniently and intuitively analyzed.

Therefore, in the embodiment, the learning condition is comprehensively displayed in the multi-target learning stage, and a user can conveniently analyze the learning condition according to the requirement.

The scheme also provides a machine learning progress control method based on the working condition, which is applied to electronic equipment of a progress control system, wherein the progress control system comprises a server and the electronic equipment, and the method comprises the following steps: sending a progress inquiry request to the server, wherein the progress inquiry request comprises a target working condition code, the target working condition code is used for indicating a target working condition, the target working condition is used for indicating a current learning condition, the current learning condition is used for indicating the current learning condition of the target working condition in each learning stage, the current learning condition is used for determining whether a preliminary learning stage exists in the target working condition, and the preliminary learning stage is a learning stage which can be entered but is not entered in the target working condition at present; if the pre-learning stage exists, acquiring the current learning condition and prompt information sent by the server, wherein the prompt information is used for prompting a user to create a task of the target working condition in the pre-learning stage; and if the preliminary learning stage does not exist, acquiring the current learning situation from the server.

Therefore, in the example, the user can obtain the machine learning condition corresponding to each working condition in real time through the electronic equipment, and the user can conveniently manage and control the whole machine learning progress.

Referring to fig. 8, fig. 8 is a block diagram illustrating functional units of a device for managing machine learning progress based on operating conditions according to an embodiment of the present disclosure. The machine learning progress control apparatus 300 based on the working condition is applied to a server of a progress control system including the server and an electronic device, and includes: an obtaining unit 301, configured to obtain a progress query request from the electronic device, where the progress query request includes a target operating condition code, and the target operating condition code is used to indicate a target operating condition; a first determining unit 302, configured to determine, according to the target operating condition code, a current learning condition of the target operating condition, where the current learning condition is used to indicate a current learning condition of the target operating condition in each learning stage; a second determining unit 303, configured to determine whether a preliminary learning stage exists in the target operating condition according to the current learning condition, where the preliminary learning stage is a learning stage that the target operating condition may currently enter but has not yet entered; a first sending unit 304, configured to send the current learning situation and prompt information to the electronic device if the pre-learning stage exists, where the prompt information is used to prompt a user to create a task of the target working condition in the pre-learning stage; a second sending unit 305, configured to send the current learning situation to the electronic device if the preliminary learning stage does not exist.

In a possible example, the learning progress of the target condition includes a learning task creating stage, a learning target confirming stage, a learning production scheduling stage, a training stage, a multi-target learning stage, and a verification and evaluation stage, and in the aspect of determining whether the target condition has a preliminary learning stage according to the current learning condition, the second determining unit 303 is specifically configured to: determining a next learning stage of the current learning stage, wherein the learning task creating stage is an initial learning stage, the learning target confirming stage is the next learning stage of the learning task creating stage, the learning production scheduling stage is the next stage of the learning target confirming stage, the training stage is the next stage of the learning production scheduling stage, the multi-target learning stage is the next stage of the training stage, and the verification and evaluation stage is the next stage of the multi-target learning stage; determining whether the learning task of the target working condition in the current learning stage is finished; and if so, determining that the next learning stage of the current learning stage is a preparatory learning stage.

In one possible example, the initial stage of the learning progress of the target condition is a stage of creating a learning task, the current learning condition of the target condition includes that the target condition has not created a learning task, and after the determining the current learning condition of the target condition according to the target condition code, the apparatus 300 is further configured to: determining the working condition type of the target working condition according to the target working condition code; determining the number of recommended tasks according to the working condition types, wherein the number of recommended tasks is used for recommending the number of learning tasks which need to be created by the user in the learning task creating stage and aim at the target working condition; and sending the recommended task quantity to the electronic equipment.

In one possible example, in the aspect of determining the recommended number of tasks according to the operating condition type, the apparatus 300 is configured to: determining a same-type working condition code according to the target working condition code, wherein the similarity between the same-type working condition code and the target working condition code is greater than a preset value; acquiring a preparation working condition, wherein the preparation working condition is a working condition that all learning stages are completed in the same working condition corresponding to the same working condition codes; determining the number of learning tasks corresponding to the preparation condition and the product comprehensive quality index corresponding to each learning task corresponding to the preparation condition; determining the quality distribution condition according to the product comprehensive quality index of each learning task; determining a recommended task quantity interval according to the quality distribution condition and the learning task quantity; and determining the recommended task quantity according to the recommended task quantity interval.

In a possible example, the current learning condition includes a learning condition of a training phase, and in the aspect of determining the current learning condition of the target condition according to the target condition code, the first determining unit 302 is specifically configured to: determining a training total batch of the target working condition in the training stage; obtaining the production condition of each production flow aiming at the target working condition, wherein the production condition is used for indicating the production condition of the current production batch corresponding to the production flow; determining a completion progress corresponding to each production flow according to the training total batch and the current production batch; determining the training progress of the target working condition in the training stage according to the completion progress corresponding to each production flow; in terms of the sending the current learning situation to the electronic device, the first sending unit 304 or the second sending unit 305 is configured to: and sending the production condition and the training progress to the electronic equipment.

In a possible example, in the aspect that the training progress of the target operating condition in the training phase is determined according to the completion progress corresponding to each production flow, the first determining unit 302 is specifically configured to: determining a preparation production flow, wherein the completion progress corresponding to the preparation production flow is less than 1; numbering the preparation production flow according to a production sequence, wherein the production sequence is more backward as the numbering sequence is larger; determining a progress coefficient for each preliminary production flow according to a production sequence, wherein the progress coefficient is higher the later the production sequence of the preliminary production flow is; and determining the training progress of the target working condition in the training stage according to the progress coefficient and the completion progress corresponding to the preparation production flow.

In a possible example, when the current learning condition includes a multi-target learning phase, the first determining unit 302 is specifically configured to determine, according to the target condition code, a current learning condition of the target condition: determining a learned turn of the target operating condition; acquiring product parameter information obtained by learning in each learning turn in the learned turns; sorting the learned turns according to the product parameter information; in terms of the sending the current learning situation to the electronic device, the first sending unit 304 or the second sending unit 305 is specifically configured to: sending product parameter information corresponding to a target learning turn to the electronic equipment, wherein the target learning turn is a learned turn sequenced before a preset sequence; after the product parameter information corresponding to the target learning turn is sent to the electronic device, the method further includes: receiving a setting request from the electronic equipment for the versions of the target learning turns, wherein the learning parameter corresponding to the multi-target learning of each turn is one version; updating the version of the target learning round to a preset version according to the version setting request; and determining the preset version as a version which can enter a verification evaluation phase.

It can be understood that, since the method embodiment and the apparatus embodiment are different presentation forms of the same technical concept, the content of the method embodiment portion in the present application should be synchronously adapted to the apparatus embodiment portion, and is not described herein again.

In the case of adopting an integrated unit, as shown in fig. 9, fig. 9 is a block diagram of functional units of another condition-based machine learning progress control apparatus provided in the embodiment of the present application. In fig. 9, the operating condition-based machine learning progress controlling apparatus 400 includes: a processing module 412 and a communication module 411. The processing module 412 is used for controlling and managing actions of the condition-based machine learning progress management device, for example, executing the steps of the acquiring unit 301, the first determining unit 302, the second determining unit 303, the first transmitting unit 304, and the second transmitting unit 305, and/or other processes for executing the techniques described herein. The communication module 411 is used for interaction between the machine learning progress control device based on the working condition and other equipment. As shown in fig. 9, the condition-based machine learning progress control apparatus may further include a storage module 413, and the storage module 413 is used for storing program codes and data of the condition-based machine learning progress control apparatus.

The Processing module 412 may be a Processor or a controller, and may be, for example, a Central Processing Unit (CPU), a general-purpose 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. Which may implement or execute the various illustrative logical blocks, modules, and circuits described in connection with the disclosure herein. The processor may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs, and microprocessors, among others. The communication module 411 may be a transceiver, an RF circuit or a communication interface, etc. The storage module 413 may be a memory.

All relevant contents of each scene related to the method embodiment may be referred to the functional description of the corresponding functional module, and are not described herein again. The condition-based machine learning progress control apparatus 400 may perform the condition-based machine learning progress control method shown in fig. 4.

Referring to fig. 10, fig. 10 is a block diagram illustrating functional units of another working condition-based machine learning progress control apparatus according to an embodiment of the present disclosure. The working condition-based machine learning progress control apparatus 500 is applied to an electronic device of a progress control system including a server and the electronic device, and includes: a sending unit 501, configured to send a progress query request to the server, where the progress query request includes a target working condition code, the target working condition code is used to indicate a target working condition, the target working condition is used to indicate a current learning condition, the current learning condition is used to indicate a learning condition of the target working condition in each learning stage, the current learning condition is used to determine whether a preliminary learning stage exists in the target working condition, and the preliminary learning stage is a learning stage that the target working condition can currently enter but does not yet enter; a first obtaining unit 502, configured to obtain the current learning condition and prompt information sent by the server if the pre-learning stage exists, where the prompt information is used to prompt a user to create a task of the target working condition in the pre-learning stage; a second obtaining unit 503, configured to obtain the current learning status from the server if the preliminary learning stage does not exist.

In the case of adopting an integrated unit, as shown in fig. 11, fig. 11 is a block diagram of functional units of another condition-based machine learning progress control apparatus provided in the embodiment of the present application. In fig. 11, the condition-based machine learning progress control apparatus 600 includes: a processing module 612 and a communication module 611. The processing module 612 is used for controlling and managing actions of the condition-based machine learning progress management device, for example, executing the steps of the sending unit 501, the first obtaining unit 502, and the second obtaining unit 503, and/or other processes for executing the techniques described herein. The communication module 611 is used for interaction between the machine learning progress control device and other devices based on the working conditions. As shown in fig. 11, the condition-based machine learning progress control apparatus may further include a storage module 613, and the storage module 613 is used for storing program codes and data of the condition-based machine learning progress control apparatus.

The Processing module 612 may be a Processor or a controller, and may be, for example, a Central Processing Unit (CPU), a general-purpose 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. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor may also be a combination of computing functions, e.g., comprising one or more microprocessors, DSPs, and microprocessors, among others. The communication module 611 may be a transceiver, an RF circuit or a communication interface, etc. The storage module 613 may be a memory.

All relevant contents of each scene related to the method embodiment may be referred to the functional description of the corresponding functional module, and are not described herein again.

The above description has introduced the solution of the embodiment of the present application mainly from the perspective of the method-side implementation process. It is understood that the electronic device includes hardware structures and software modules for performing the respective functions in order to realize the functions. Those of skill in the art will readily appreciate that the present application is capable of hardware or a combination of hardware and computer software implementing the various illustrative elements and algorithm steps described in connection with the embodiments provided herein. Whether a function is performed as hardware or computer software drives hardware depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

In the embodiment of the present application, the electronic device may be divided into the functional units according to the method example, for example, each functional unit may be divided corresponding to each function, or two or more functions may be integrated into one processing unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit. It should be noted that the division of the unit in the embodiment of the present application is schematic, and is only a logic function division, and there may be another division manner in actual implementation.

Embodiments of the present application further provide a chip, where the chip includes a processor, configured to call and run a computer program from a memory, so that a device in which the chip is installed performs some or all of the steps described in the electronic device in the above method embodiments.

Embodiments of the present application further provide a computer storage medium, where the computer storage medium stores a computer program for electronic data exchange, the computer program enables a computer to execute part or all of the steps of any one of the methods as described in the above method embodiments, and the computer includes an electronic device.

Embodiments of the present application also provide a computer program product comprising a non-transitory computer readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods as described in the above method embodiments. The computer program product may be a software installation package, the computer comprising an electronic device.

It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present application is not limited by the order of acts described, as some steps may occur in other orders or concurrently depending on the application. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required in this application.

In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus may be implemented in other manners. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the above-described division of the units is only one type of division of logical functions, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of some interfaces, devices or units, and may be an electric or other form.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit may be stored in a computer readable memory if it is implemented in the form of a software functional unit and sold or used as a stand-alone product. Based on such understanding, the technical solution of the present application may be substantially implemented or a part of or all or part of the technical solution contributing to the prior art may be embodied in the form of a software product stored in a memory, and including several instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the above-mentioned method of the embodiments of the present application. And the aforementioned memory comprises: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable memory, which may include: flash Memory disks, Read-Only memories (ROMs), Random Access Memories (RAMs), magnetic or optical disks, and the like.

The foregoing detailed description of the embodiments of the present application has been presented to illustrate the principles and implementations of the present application, and the above description of the embodiments is only provided to help understand the method and the core concept of the present application; meanwhile, for a person skilled in the art, according to the idea of the present application, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present application.

Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications can be easily made by those skilled in the art without departing from the spirit and scope of the present invention, and it is within the scope of the present invention to include different functions, combination of implementation steps, software and hardware implementations.

Claims

1. A machine learning progress control method based on working conditions is characterized by comprising the following steps:

acquiring a progress query request from electronic equipment, wherein the progress query request comprises a target working condition code, and the target working condition code is used for indicating a target working condition;

2. The method of claim 1, wherein the learning progress of the target condition comprises a learning task creating stage, a learning target confirming stage, a learning production scheduling stage, a training stage, a multi-target learning stage and a verification and evaluation stage, and the determining whether the target condition has a preliminary learning stage according to the current learning condition comprises:

determining a next learning stage of the current learning stage, wherein the learning task creating stage is an initial learning stage, the learning target confirming stage is the next learning stage of the learning task creating stage, the learning production scheduling stage is the next stage of the learning target confirming stage, the training stage is the next stage of the learning production scheduling stage, the multi-target learning stage is the next stage of the training stage, and the verification and evaluation stage is the next stage of the multi-target learning stage;

determining whether the learning task of the target working condition in the current learning stage is finished;

and if so, determining that the next learning stage of the current learning stage is a preparatory learning stage.

3. The method of claim 1 or 2, wherein an initial stage of the learning progress of the target condition is a learning task creation stage, the current learning condition of the target condition comprises that the target condition has not created a learning task, and after determining the current learning condition of the target condition according to the target condition code, the method further comprises:

determining the working condition type of the target working condition according to the target working condition code;

determining the number of recommended tasks according to the working condition types, wherein the number of recommended tasks is used for recommending the number of learning tasks which need to be created by the user in the learning task creating stage and aim at the target working condition;

and sending the recommended task quantity to the electronic equipment.

4. The method of claim 3, wherein determining the recommended number of tasks based on the type of condition comprises:

determining the same-type working condition codes according to the target working condition codes, wherein the similarity between the same-type working condition codes and the target working condition codes is greater than a preset value;

acquiring a preparation working condition, wherein the preparation working condition is a working condition that all learning stages are completed in the same working condition corresponding to the same working condition codes;

determining the number of learning tasks corresponding to the preparation condition and the product comprehensive quality index corresponding to each learning task corresponding to the preparation condition;

determining the quality distribution condition according to the product comprehensive quality index of each learning task;

determining a recommended task quantity interval according to the quality distribution condition and the learning task quantity;

and determining the recommended task quantity according to the recommended task quantity interval.

5. The method of any one of claims 1 to 4, wherein the current learning condition comprises a learning condition of a training phase, and the determining the current learning condition of the target condition according to the target condition code comprises:

determining a training total batch of the target working condition in the training stage;

obtaining the production condition of each production flow aiming at the target working condition, wherein the production condition is used for indicating the production condition of the current production batch corresponding to the production flow;

determining a completion progress corresponding to each production flow according to the training total batch and the current production batch;

determining the training progress of the target working condition in the training stage according to the completion progress corresponding to each production flow;

the sending the current learning situation to the electronic device includes:

and sending the production condition and the training progress to the electronic equipment.

6. A machine learning progress control method based on working conditions is characterized by comprising the following steps:

sending a progress inquiry request to a server, wherein the progress inquiry request comprises a target working condition code, the target working condition code is used for indicating a target working condition, the target working condition is used for indicating a current learning condition, the current learning condition is used for indicating the current learning condition of the target working condition in each learning stage, the current learning condition is used for determining whether a preliminary learning stage exists in the target working condition, and the preliminary learning stage is a learning stage which can be entered but is not entered in the target working condition at present;

7. The utility model provides a machine learning progress management and control device based on operating mode which characterized in that, the device includes:

the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring a progress query request from electronic equipment, the progress query request comprises a target working condition code, and the target working condition code is used for indicating a target working condition;

a second sending unit, configured to send the current learning situation to the electronic device if the pre-learning stage does not exist.

8. The utility model provides a machine learning progress management and control device based on operating mode which characterized in that, the device includes:

a sending unit, configured to send a progress query request to a server, where the progress query request includes a target condition code, the target condition code is used to indicate a target condition, the target condition is used to indicate a current learning situation, the current learning situation is used to indicate a learning situation of the target condition in each learning stage, and the current learning situation is used to determine whether a pre-learning stage exists in the target condition, where the pre-learning stage is a learning stage that the target condition may enter but does not enter currently;

9. An apparatus comprising a processor, a memory, and one or more programs stored in the memory and configured to be executed by the processor, the programs comprising instructions for performing the steps in the method of any of claims 1-6.

10. A computer-readable storage medium or computer program product, characterized in that the computer-readable storage medium stores a computer program, the computer program causing a computer to perform the method of any of claims 1-6, or the computer program product causing a computer to perform the method of any of claims 1-6.