Method and device for executing data enhancement operation, and computer equipment

Abstract

The embodiments of the present application provide a method and device for executing a data enhancement operation, and a computer device, which belong to the field of data enhancement technology. The method includes: obtaining at least one data enhancement operation to be executed and the execution strategy selection information of each of the data enhancement operations; obtaining the number of the data enhancement operations to obtain the number of operations; screening at least one target execution strategy from the preset candidate execution strategies according to the number of operations and the execution strategy selection information to obtain a target execution strategy set; executing at least one of the data enhancement operations according to the target execution strategy set. The embodiments of the present application can improve the execution efficiency of data enhancement operations and expand the versatility of data enhancement operations.

Description

Technical Field

[0001] The present application relates to the technical field of data enhancement, and in particular to a method and apparatus for executing a data enhancement operation, and a computer device. Background Art

[0002] At present, data augmentation operations can increase the quantity and diversity of data, so that algorithm developers can use the data sets after data augmentation operations to improve model performance. However, there may be conflicts between different data augmentation operations, or they may need to cooperate with each other. In related technologies, the execution methods of data augmentation operations are either executed in sequence or selected one by one, resulting in a single execution method of data augmentation operations, low versatility, and low execution efficiency of data augmentation operations. Therefore, how to improve the execution efficiency of data augmentation operations and improve the versatility of data augmentation operations has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a method and apparatus for executing data enhancement operations, and a computer device, aiming to improve the execution efficiency of data enhancement operations and expand the versatility of data enhancement operations.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application proposes a method for performing a data enhancement operation, the method comprising:

[0005] Acquire at least one data enhancement operation to be performed and execution strategy selection information of each of the data enhancement operations;

[0006] Obtaining the number of the data enhancement operations to obtain the number of operations;

[0007] Filtering at least one target execution strategy from preset candidate execution strategies according to the number of operations and the execution strategy selection information to obtain a target execution strategy set;

[0008] At least one of the data enhancement operations is performed according to the target execution strategy set.

[0009] In some embodiments, the candidate execution strategies include a single execution strategy and a composite execution strategy, the single execution strategy includes: a random interval execution strategy and a random execution strategy; the composite execution strategy includes: a combined execution strategy, an out-of-order execution strategy, a selective execution strategy and a partial execution strategy; the selecting a target execution strategy set from the preset candidate execution strategies according to the number of operations and the execution strategy selection information includes:

[0010] If the number of operations is equal to 1, the target execution strategy set is selected from the random interval execution strategy and the random execution strategy according to the execution strategy selection information;

[0011] If the number of operations is greater than 1, the target execution strategy set is selected from the combined execution strategy, the out-of-order execution strategy, the alternative execution strategy and the partial execution strategy according to the execution strategy selection information.

[0012] In some embodiments, if the target execution strategy set is the random interval execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes:

[0013] Obtaining at least one group of reference intervals for the data enhancement operation;

[0014] generating a first random number for each data enhancement operation according to the reference interval;

[0015] The formal parameter of the data enhancement operation is used as a key and the first random number is used as a value to store in a preset parameter list to obtain a first parameter list;

[0016] The data enhancement operation is performed according to the first parameter list.

[0017] In some embodiments, if the target execution strategy set is the random execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes:

[0018] Generate a second random number according to the data augmentation operation;

[0019] Compare the second random number with a preset threshold to obtain a comparison result;

[0020] The data enhancement operation is performed according to the comparison result.

[0021] In some embodiments, if the target execution strategy includes the combined execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes:

[0022] Obtaining a sorting sequence number for each of the data enhancement operations;

[0023] Each of the data enhancement operations is performed in sequence according to the sorting sequence number.

[0024] In some embodiments, if the target execution strategy set includes: an out-of-order execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes:

[0025] Generate an execution sequence number list according to at least one of the data enhancement operations;

[0026] Extracting the current execution sequence number from the execution sequence number list;

[0027] If the current execution sequence number is not empty, perform the data enhancement operation according to the current execution sequence number, and shuffle the sequence numbers in the execution sequence number list;

[0028] Obtain the next execution sequence number of the scrambled execution sequence number list, and if the next execution sequence number is not empty, perform the next data enhancement operation according to the next execution sequence number; if the next execution sequence number is empty, skip it.

[0029] In some embodiments, if the target execution strategy set includes: selecting an execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes:

[0030] Randomly extracting a selected enhancement operation from at least one of the data enhancement operations;

[0031] The selected enhancement operation is performed.

[0032] In some embodiments, if the target execution strategy set includes: a partial execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes:

[0033] Obtaining a threshold value for each of the data enhancement operations to obtain a reference threshold value;

[0034] Generate a third random number according to the data augmentation operation;

[0035] If the third random number is less than the reference threshold, the data enhancement operation is performed.

[0036] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a device for performing a data enhancement operation, the device comprising:

[0037] A data acquisition module, used to acquire at least one data enhancement operation to be performed and execution strategy selection information of each of the data enhancement operations;

[0038] A quantity acquisition module, used to acquire the quantity of the data enhancement operations to obtain the operation quantity;

[0039] A policy screening module, configured to screen out at least one target execution policy from preset candidate execution policies according to the number of operations and the execution policy selection information, and obtain a target execution policy set;

[0040] An operation execution module is used to execute at least one of the data enhancement operations according to the target execution strategy set.

[0041] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0042] The method and device for executing data enhancement operations and the computer equipment proposed in this application select a target execution strategy set from multiple candidate execution strategies by determining the number of data enhancement operations and execution strategy selection information, so as to execute the corresponding data enhancement operations one by one through each target execution strategy in the target execution strategy set, so as to be applicable to different data enhancement operation execution scenarios and improve the efficiency of data enhancement. At the same time, when implementing the execution of data enhancement operations, the execution of data enhancement operations can be quickly completed by directly selecting available execution strategies, thereby improving the execution efficiency of data enhancement operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is a flowchart of a method for executing a data enhancement operation provided in an embodiment of the present application;

[0044] Figure 2 is a schematic diagram of a data enhancement preprocessing framework in a method for executing a data enhancement operation provided in an embodiment of the present application;

[0045] Figure 3 yes Figure 1 Flow chart of step S103 in FIG.

[0046] Figure 4 yes Figure 1 Flow chart of step S104 in FIG.

[0047] Figure 5 yes Figure 1 Another flowchart of step S104 in;

[0048] Figure 6 yes Figure 1 Another flowchart of step S104 in;

[0049] Figure 7 yes Figure 1 Another flowchart of step S104 in;

[0050] Figure 8 yes Figure 1 Another flowchart of step S104 in;

[0051] Fig. 9 yes Figure 1 Another flowchart of step S104 in;

[0052] Fig.10 yes Figure 1Another flowchart of step S104 in;

[0053] Fig.11 is a schematic diagram of the structure of a device for executing a data enhancement operation provided in an embodiment of the present application;

[0054] Fig.12 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0057] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0058] First, some nouns involved in this application are analyzed:

[0059] Data Augmentation: Also known as data amplification, it means making limited data produce the equivalent value of more data without substantially increasing the data, and increasing the amount and diversity of data by enhancing the original data.

[0060] AI: artificial intelligence.

[0061] Data enhancement operation: It can be understood as a function of a specific function. A complete data enhancement module is often expressed as a series of operation combinations.

[0062] By enhancing the original data, the original data can be transformed into data of different styles to expand the amount of data. Data enhancement is an indispensable step in the model training process, which enables the model to learn more data features to help enhance the generalization function of the model. For example, in image data, the main ways to enhance image data are random cropping, offset, scaling, rotation, distortion, and changes in color and brightness, and even adding some noise, occlusion and other operations to the image, the purpose is to increase the amount and diversity of data to improve the performance of the model. However, in the process of executing data enhancement operations, some data enhancement operations will conflict with each other, some data enhancement operations need to cooperate with each other, and some data enhancement operations do not interfere with each other and run in parallel. Moreover, in the related art, data enhancement operations are generally performed in sequence, and it is impossible to adjust the execution of data enhancement operations according to different needs, which makes the function of data enhancement operations single and the execution efficiency of data enhancement operations low.

[0063] Based on this, the embodiments of the present application provide a method and apparatus for executing data enhancement operations, and a computer device, which is intended to select a target execution strategy set from multiple candidate execution strategies by determining the number of data enhancement operations and execution strategy selection information, so as to execute the corresponding data enhancement operations one by one through each target execution strategy in the target execution strategy set, so as to be applicable to different data enhancement operation execution scenarios and improve the efficiency of data enhancement. At the same time, when implementing the execution of data enhancement operations, directly selecting an available execution strategy can quickly complete the execution of data enhancement operations, thereby improving the execution efficiency of data enhancement operations.

[0064] The execution method, apparatus, and computer device of the data enhancement operation provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the execution method of the data enhancement operation in the embodiments of the present application is described.

[0065] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0066] The execution method of the data enhancement operation provided in the embodiment of the present application relates to the field of data enhancement technology. The execution method of the data enhancement operation provided in the embodiment of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application of the execution method of the data enhancement operation, etc., but is not limited to the above forms.

[0067] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0068] Figure 1 is an optional flowchart of a method for executing a data enhancement operation provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S104.

[0069] Step S101, obtaining at least one data enhancement operation to be performed and execution strategy selection information of each data enhancement operation;

[0070] Step S102, obtaining the number of data enhancement operations to obtain the number of operations;

[0071] Step S103, selecting at least one target execution strategy from preset candidate execution strategies according to the number of operations and the execution strategy selection information, and obtaining a target execution strategy set;

[0072] Step S104: Execute at least one data enhancement operation according to the target execution strategy set.

[0073] In the steps S101 to S104 shown in the embodiment of the present application, after determining the execution strategy selection information and the total number of operations of each data enhancement operation, the target execution strategy set of each data enhancement operation is screened out according to the execution strategy selection information and the number of operations, so as to perform at least one data enhancement operation according to the target execution strategy set. By setting a plurality of different types of execution strategies for selection, different data enhancement operations can be performed according to different types of execution strategies, which can not only improve the versatility of the execution of the data enhancement operation, but also improve the execution efficiency of the data enhancement operation.

[0074] In step S101 of some embodiments, the data enhancement operation is a specific function or function, and the data enhancement operation is different in different fields. For example, in the field of images, the data enhancement operation can be random cropping, offset, scaling, rotation, distortion, noise addition, occlusion, and illumination color change, etc. In the field of speech, the data enhancement operation can be time domain enhancement, frequency domain enhancement, and multi-domain enhancement, etc. In the field of text, data enhancement can be word replacement, word deletion, synonym replacement, and translation, etc. Therefore, in different fields, the data enhancement operation is different, so in order to be applicable to data enhancement in multiple fields, that is, to improve the versatility of data enhancement, this embodiment selects corresponding execution strategies for different data enhancement operations to achieve enhanced processing of data in different fields. The execution strategy selection information represents the selection information for executing at least one data enhancement operation, and the execution strategy selection information is mainly determined according to the category and association information of the data enhancement operation. It should be noted that the association information represents whether there is a conflict between different data enhancement operations, whether cooperation is required, and whether interference is involved, etc., so as to determine in which way the data in different fields are enhanced according to different data enhancement operations, quickly construct a large amount of data, and improve the efficiency of model optimization.

[0075] In step S102 of some embodiments, Figure 2 As shown, this embodiment sets up a data enhancement preprocessing framework applied to various fields. The data enhancement preprocessing framework is divided into a single execution strategy and a composite execution strategy. The single execution strategy is used for the random execution of a single data enhancement operation, while the composite execution operation is used for the execution of multiple data enhancements. Therefore, by obtaining the number of data enhancement operations, it is further determined which execution strategy to select according to the number of operations.

[0076] See also Figure 3In some embodiments, the candidate execution strategies include a single execution strategy and a composite execution strategy. The single execution strategy includes: a random interval execution strategy and a random execution strategy; the composite execution strategy includes: a combined execution strategy, an out-of-order execution strategy, a selective execution strategy and a partial execution strategy; step S103 may include but is not limited to steps S301 to S302:

[0077] Step S301, if the number of operations is equal to 1, a target execution strategy set is selected from the random interval execution strategy and the random execution strategy according to the execution strategy selection information;

[0078] Step S302: If the number of operations is greater than 1, a target execution strategy set is selected from a combination execution strategy, an out-of-order execution strategy, a selective execution strategy, and a partial execution strategy according to the execution strategy selection information.

[0079] In step S301 of some embodiments, if the number of operations is equal to 1, it indicates that only a single data enhancement operation needs to be performed, so the execution strategy with the composite requirement is screened out from the single execution strategy. It should be noted that the single execution strategy includes: a random execution strategy and a random interval execution strategy, so the target execution strategy set is directly screened out from the random interval execution strategy and the random execution strategy according to the execution strategy selection information of the data enhancement operation, so as to determine whether the random interval execution strategy or the random execution strategy is used as the target execution strategy in the target execution strategy set.

[0080] In step S302 of some embodiments, if the number of operations is greater than 1, that is, there are multiple data enhancement operations to be processed, the execution strategy of selecting multiple data enhancement operations in the composite execution operation is located. It should be noted that the composite execution operation includes: a combined execution strategy, an out-of-order execution strategy, a selected execution strategy, and a partial execution strategy, so at least one execution strategy is directly selected from the combined execution strategy, the out-of-order execution strategy, the selected execution strategy, and the partial execution strategy to construct a target execution strategy set, so that multiple different data enhancement operations are executed with different execution strategies.

[0081] In steps S301 to S302 shown in this embodiment, different execution strategies are selected according to the number of data enhancement operations to quickly construct an execution strategy set for each data enhancement operation.

[0082] In step S104 of some embodiments, the target execution strategy set may include at least one of: a random interval execution strategy, a random execution strategy, a combined execution strategy, an out-of-order execution strategy, a selective execution strategy and a partial execution strategy, so as to be suitable for the execution methods of data enhancement operations in different fields.

[0083] See also Figure 4In some embodiments, if the target execution strategy set includes only one target execution strategy, and the target execution strategy is a random execution strategy, step S104 may include but is not limited to steps S401 to S403:

[0084] Step S401, generating a second random number according to the data enhancement operation;

[0085] Step S402, comparing the second random number with a preset threshold to obtain a comparison result;

[0086] Step S403: performing a data enhancement operation according to the comparison result.

[0087] In some embodiments, in steps S401 to S402, when the target execution strategy is a random execution strategy, that is, the data enhancement operation is performed in a random execution manner, but whether to perform the operation still needs to be determined according to the random number of the data enhancement operation. Therefore, a random number is randomly set for the data enhancement operation as the second random number, and the second random number is compared with the preset threshold to determine whether the data enhancement operation needs to be performed, thereby realizing the random execution strategy of the data enhancement operation.

[0088] In step S403 of some embodiments, a data enhancement operation is performed according to a comparison result between the second random number and a preset threshold. It should be noted that if the second random number exceeds the preset threshold, the data enhancement operation is not performed, and if the second random number does not exceed the preset threshold, the data enhancement operation is performed. Therefore, by determining whether the data enhancement operation is performed based on the comparison result between the second random number and the preset threshold, the random execution of the data enhancement operation is achieved.

[0089] In steps S401 to S403 shown in this embodiment, by setting a second random number for the data enhancement operation and judging whether to perform the data enhancement operation according to the comparison result between the second random number and a preset threshold, the random execution of the data enhancement operation is achieved, thereby meeting the development requirements of randomly executing the data enhancement operation.

[0090] See also Figure 5 In some embodiments, if the target execution strategy set includes only one target execution strategy, and the target execution strategy is a random interval execution strategy, step S104 may include but is not limited to steps S501 to S504:

[0091] Step S501, obtaining at least one group of reference intervals for data enhancement operations;

[0092] Step S502, generating a first random number for each data enhancement operation according to the reference interval;

[0093] Step S503, using the parameter of the data enhancement operation as a key and the first random number as a value to store in a preset parameter list, to obtain a first parameter list;

[0094] Step S504: performing a data enhancement operation according to the first parameter list.

[0095] In step S501 of some embodiments, if the data enhancement operation needs to be executed with a random interval execution strategy, at least one group of reference intervals for the data enhancement operation, that is, N groups of intervals, is first determined. It should be noted that the reference interval is the execution interval of the data enhancement operation, and the data enhancement operation can only be executed when the random number is within the reference interval, thereby realizing the random interval execution of the data enhancement operation.

[0096] In step S502 of some embodiments, the first random number is a random number used to characterize the data enhancement operation and is a reference value for determining whether the data enhancement operation is performed. It should be noted that each time the data enhancement operation is performed, N first random numbers are generated according to the N groups of intervals, and these first random numbers are passed to the enhancement function corresponding to the data enhancement operation.

[0097] In some embodiments, in steps S503 to S504, after the first random number of each data augmentation operation is determined, the formal parameter of the data augmentation operation is used as a key, and the first random number is stored as a value in a preset parameter list to obtain a first parameter list. Therefore, after the first parameter list of the data augmentation operation is completed, the completed data augmentation operation is performed according to the first parameter list, so that the data augmentation operation is completely executed.

[0098] In steps S501 to S504 shown in this embodiment, when it is necessary to perform a data enhancement operation in a random interval, it is necessary to first determine the first random number of each data enhancement operation. If it is determined that the first random number is within the reference interval, the data enhancement operation is performed to achieve random interval execution of the data enhancement operation, so as to meet the developer's requirements for random execution of the data enhancement operation.

[0099] It should be noted that, in the data augmentation operation, the random execution strategy and the random interval execution strategy can be merged so that the data augmentation operation can be executed with two execution strategies through the merged manner.

[0100] See also Figure 6 In some embodiments, if the target execution policy set includes at least one target execution policy, and the target execution policy is a combined execution policy, step S104 may include but is not limited to steps S601 to S602:

[0101] Step S601, obtaining the sorting sequence number of each data enhancement operation;

[0102] Step S602, executing each data enhancement operation in sequence according to the sorting sequence number.

[0103] In step S601 of some embodiments, when it is necessary to perform a data enhancement operation according to a combined execution strategy, the combined execution strategy can also be called sequential execution, and each data enhancement operation is performed in sequence. It should be noted that the data enhancement module includes a series of data enhancement operations. For example, the data enhancement module of image data can be a combination of cropping, rotation and color change, but whether all or part of these data enhancement operations need to be executed depends on the needs of the developer. If the combined execution strategy is used, it means that each data enhancement operation needs to be executed in sequence, so it is necessary to first obtain the sequence number of each data enhancement operation as the sorting sequence number.

[0104] In step S602 of some embodiments, each time a data enhancement operation is performed, the data enhancement operations are performed one by one according to the order of the sorting sequence numbers from small to large, so as to achieve sequential execution of at least one data enhancement operation.

[0105] In steps S601 to S602 shown in this embodiment, when the data enhancement operation needs to be performed in a combined execution manner, the serial number of each data enhancement operation is first determined, and then the operations are performed one by one according to the serial number, so that the data enhancement operations are executed one by one, meeting the needs of developers to execute the data enhancement operations one by one.

[0106] See also Figure 7 In some embodiments, if the target execution strategy set includes at least one target execution strategy, and the target execution strategy is an out-of-order execution strategy, step S104 may include but is not limited to steps S701 to S704:

[0107] Step S701, generating an execution sequence number list according to at least one data enhancement operation;

[0108] Step S702, extracting the current execution sequence number from the execution sequence number list;

[0109] Step S703, if the current execution sequence number is not empty, perform data enhancement operation according to the current execution sequence number, and shuffle the sequence numbers in the execution sequence number list;

[0110] Step S704, obtaining the next execution sequence number of the scrambled execution sequence number list, if the next execution sequence number is not empty, performing the next data enhancement operation according to the next execution sequence number, if the next execution sequence number is empty, skipping it.

[0111] In step S701 of some embodiments, if the data enhancement operation needs to be executed in a disordered manner, it is necessary to first construct an execution sequence number list according to the data enhancement operation. It should be noted that the execution sequence number list is determined by the execution order of the data enhancement operation and is used to store the sequence number corresponding to each data enhancement operation, but the sequence numbers in the execution sequence number list are not sorted in order, but randomly arranged.

[0112] In steps S702 to S703 of some embodiments, when a data enhancement operation needs to be performed, the current execution sequence number is first extracted from the execution sequence number list. If the data enhancement operation corresponding to the current execution sequence number is empty, it is directly skipped. If the data enhancement operation corresponding to the current execution sequence number is not empty, the corresponding data enhancement operation is performed according to the current execution sequence number. It should be noted that in order to prevent interference between the execution of each data enhancement operation, the sequence numbers in the execution sort list will be shuffled after each execution of the data enhancement operation, and the shuffling method is random.

[0113] In step S704 of some embodiments, the next execution sequence number is obtained in the shuffled execution sequence number list in order to execute the corresponding data enhancement operation according to the next execution sequence number. It should be noted that after completing the data enhancement operation, the execution sequence number list continues to be shuffled, and then the data enhancement operation continues to be executed, that is, steps S702 to S704 are repeatedly executed, so that after each data enhancement operation is executed, the execution sequence number list is shuffled and then the execution of the data enhancement operation is decided in sequence. For example, if the data enhancement operation with the execution sequence number k is determined when the data enhancement operation is executed for the nth time, the data enhancement operation with the execution sequence number k-6 may be executed for the n+1th time, so as to realize the out-of-order execution of the data enhancement operation.

[0114] In steps S701 to S704 shown in this embodiment, when the data enhancement operation needs to be executed with an out-of-order execution strategy, the execution sequence number list is shuffled directly after each execution of the data enhancement operation, and then the data enhancement operation is executed according to the shuffled execution sequence number list, thereby meeting the developer's requirements for out-of-order execution of the data enhancement operation.

[0115] See also Figure 8 In some embodiments, if the target execution policy set includes at least one target execution policy, and the target execution policy is a selective execution policy, step S104 may include but is not limited to steps S801 to S802:

[0116] Step S801, randomly extracting a selected enhancement operation from at least one data enhancement operation;

[0117] Step S802, performing the selected enhancement operation.

[0118] In steps S801 to S802 of some embodiments, when it is necessary to perform a data enhancement operation according to a selective execution strategy, each time a data enhancement operation is performed, a data enhancement operation is randomly selected from at least one data enhancement operation as a selected enhancement operation, and the selected enhancement operation is performed. It should be noted that if there are N data enhancement operations, the selected enhancement operation is selected from the N data enhancement operations, and if the selected enhancement operation is an empty operation, it is directly skipped, and if the selected enhancement operation is a non-empty operation, the selected enhancement operation is performed.

[0119] In step S801 to step S802 shown in this embodiment, one data enhancement operation is randomly selected to be executed, so as to realize the selective execution of the data enhancement operations, thereby meeting the requirement of selective execution.

[0120] See also Fig. 9 In some embodiments, if the target execution policy set includes at least one target execution policy, and the target execution policy is a partial execution policy, step S104 may include but is not limited to steps S901 to S903:

[0121] Step S901, obtaining a threshold value for each data enhancement operation to obtain a reference threshold value;

[0122] Step S902, generating a third random number according to the data enhancement operation;

[0123] Step S903: If the third random number is less than the reference threshold, perform a data enhancement operation.

[0124] In step S901 of some embodiments, if the data enhancement operation needs to be executed according to a partial execution strategy, it indicates that only some of the data enhancement operations need to be executed, not all of the data enhancement operations need to be executed. It should be noted that the reference threshold is used to determine which data enhancement operations need to be executed, and each data enhancement operation sets a reference threshold, and the reference threshold is between 0 and 1.

[0125] In step S902 of some embodiments, each time a data enhancement operation is performed, a third random number is randomly generated for the data enhancement operation. The third random number is used to determine whether the data enhancement operation needs to be performed. In this embodiment, the third random number is between 0 and 1.

[0126] In step S903 of some embodiments, after traversing at least one data enhancement operation in sequence, the data enhancement operation currently being performed is determined. If the third random number corresponding to the data enhancement operation is greater than or equal to a reference threshold, the data enhancement operation is not performed. If the third random number corresponding to the data enhancement operation is less than the reference threshold, the data enhancement operation is performed to achieve partial execution of the data enhancement operation.

[0127] It should be noted that when performing a data enhancement operation, if the third random number corresponding to the data enhancement operation is less than the reference threshold, and the currently performed data enhancement operation is determined to be a null operation, it is skipped.

[0128] In steps S901 to S903 shown in this embodiment, the third random number corresponding to each data enhancement operation is determined and compared with the reference threshold to determine whether to perform the data enhancement operation, thereby achieving partial execution of the data enhancement operation to meet the need to partially execute the data enhancement operation.

[0129] See also Fig.10 In some embodiments, if the target execution policy set includes at least one target execution policy, and the target execution policy is another partial execution policy, step S104 may include but is not limited to steps S1001 to 1003:

[0130] Step S1001, obtaining the probability value of each data enhancement operation to obtain a reference probability value;

[0131] Step S1002, selecting a selected enhancement operation from the data enhancement operations according to a preset number and a reference probability value;

[0132] Step S1003, executing the selected enhancement operations in sequence.

[0133] In steps S1001 to S1003 of some embodiments, the preset number is the number of data enhancement operations that need to be performed, and the preset number is less than the number of operations to implement the execution of some data enhancement operations. The reference probability value is the probability value corresponding to each data enhancement operation, and the reference probability values ​​corresponding to different data enhancement operations are the same or different. It should be noted that the data enhancement operations are sorted according to the reference probability value, and then the preset number of data enhancement operations before sorting are used as selected data enhancements to implement partial execution of the data enhancement operations.

[0134] For example, if the reference probability value is p and the preset number is K, the first K data enhancement operations are taken as selected enhancement operations, and the selected enhancement operations are executed one by one according to the reference probability, or the selected enhancement operations are executed randomly.

[0135] In steps S1001 to S1003 shown in this embodiment, when it is necessary to select some data enhancement operations to be executed, some data enhancement operations are selected by using probability values ​​and preset quantities, and then some data enhancement operations are executed to achieve partial execution of data enhancement operations, thereby meeting the needs of different developers for the execution of data enhancement operations.

[0136] In summary, this embodiment sets a random interval execution strategy, a random execution strategy, a combined execution strategy, an out-of-order execution strategy, a select execution strategy and a partial execution strategy for different data enhancement operations to be selected, so as to meet the needs of different developers for data enhancement. In this embodiment, the data enhancement operation is positioned to the selection of each execution strategy to adapt to the enhancement of different data in different application scenarios to meet different categories of data enhancement operations, thereby improving the universal rate of data enhancement. When performing data enhancement, it is sufficient to directly filter out the execution strategy with composite requirements, so that the data enhancement operation can be efficiently executed according to the selected execution strategy, improve the efficiency of data enhancement, and then improve the data construction efficiency of the model optimization process.

[0137] See also Fig.11 The embodiment of the present application further provides a device for executing a data enhancement operation, which can implement the above-mentioned method for executing the data enhancement operation, and the device includes:

[0138] A data acquisition module 1101 is used to acquire at least one data enhancement operation to be performed and execution strategy selection information of each data enhancement operation;

[0139] A quantity acquisition module 1102 is used to acquire the quantity of data enhancement operations and obtain the operation quantity;

[0140] The policy screening module 1103 is used to screen out at least one target execution policy from preset candidate execution policies according to the number of operations and the execution policy selection information, and obtain a target execution policy set;

[0141] The operation execution module 1104 is used to execute at least one data enhancement operation according to the target execution strategy set.

[0142] The specific implementation of the device for executing the data enhancement operation is basically the same as the specific implementation of the method for executing the data enhancement operation described above, and will not be described in detail here.

[0143] The embodiment of the present application also provides a computer device, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the execution method of the above data enhancement operation when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0144] See also Fig.12 , Fig.12 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0145] The processor 1201 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0146] The memory 1202 may be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1202 may store an operating system and other applications. When the technical solution provided in the embodiments of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1202, and the processor 1201 calls and executes the execution method of the data enhancement operation in the embodiments of this application;

[0147] Input / output interface 1203, used to implement information input and output;

[0148] The communication interface 1204 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0149] A bus 1205 that transmits information between various components of the device (e.g., the processor 1201, the memory 1202, the input / output interface 1203, and the communication interface 1204);

[0150] The processor 1201 , the memory 1202 , the input / output interface 1203 and the communication interface 1204 are connected to each other in communication within the device via the bus 1205 .

[0151] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the execution method of the above-mentioned data enhancement operation is implemented.

[0152] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0153] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0154] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0155] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0156] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0157] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0158] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0159] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as 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.

[0160] The units described above 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.

[0161] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically 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 software functional units.

[0162] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0163] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A method for performing a data enhancement operation, characterized in that: The method comprises: Acquire at least one data enhancement operation to be performed and execution strategy selection information of each of the data enhancement operations; Obtaining the number of the data enhancement operations to obtain the number of operations; Filtering at least one target execution strategy from preset candidate execution strategies according to the number of operations and the execution strategy selection information to obtain a target execution strategy set; Execute at least one of the data enhancement operations according to the target execution strategy set; wherein the data of the data enhancement operation is any one of the following: image data, voice data and text data; The candidate execution strategies include a single execution strategy and a composite execution strategy. The single execution strategy includes: a random interval execution strategy and a random execution strategy; the composite execution strategy includes: a combined execution strategy, an out-of-order execution strategy, a selective execution strategy and a partial execution strategy; the target execution strategy set is selected from the preset candidate execution strategies according to the number of operations and the execution strategy selection information, including: If the number of operations is equal to 1, the target execution strategy set is selected from the random interval execution strategy and the random execution strategy according to the execution strategy selection information; If the number of operations is greater than 1, the target execution strategy set is selected from the combined execution strategy, the out-of-order execution strategy, the alternative execution strategy and the partial execution strategy according to the execution strategy selection information; If the target execution strategy set is the random interval execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes: Obtaining at least one group of reference intervals for the data enhancement operation; generating a first random number for each data enhancement operation according to the reference interval; The formal parameter of the data enhancement operation is used as a key and the first random number is used as a value to store in a preset parameter list to obtain a first parameter list; Perform the data enhancement operation according to the first parameter list; If the target execution strategy set is the random execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes: Generate a second random number according to the data augmentation operation; Compare the second random number with a preset threshold to obtain a comparison result; Perform the data enhancement operation according to the comparison result; If the target execution strategy includes the combined execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes: Obtaining a sorting sequence number for each of the data enhancement operations; Each of the data enhancement operations is performed in sequence according to the sorting sequence number.

2. The method according to claim 1, characterized in that If the target execution strategy set includes: selecting an execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes: Randomly extracting a selected enhancement operation from at least one of the data enhancement operations; The selected enhancement operation is performed.

3. The method according to any one of claims 1 to 2, characterized in that: If the target execution strategy set includes: a partial execution strategy, performing at least one of the data enhancement operations according to the target execution strategy set includes: Obtaining a threshold value for each of the data enhancement operations to obtain a reference threshold value; Generate a third random number according to the data augmentation operation; If the third random number is less than the reference threshold, the data enhancement operation is performed.

4. A device for executing a data enhancement operation, characterized in that: A method for performing a data enhancement operation according to any one of claims 1 to 3, the device comprising: A data acquisition module, used to acquire at least one data enhancement operation to be performed and execution strategy selection information of each of the data enhancement operations; A quantity acquisition module, used to acquire the quantity of the data enhancement operations to obtain the operation quantity; A policy screening module, configured to screen out at least one target execution policy from preset candidate execution policies according to the number of operations and the execution policy selection information, and obtain a target execution policy set; An operation execution module is used to execute at least one of the data enhancement operations according to the target execution strategy set.

5. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the method for executing the data enhancement operation according to any one of claims 1 to 3 when executing the computer program.

Images