Data processing method and device, electronic equipment and storage medium

Abstract

The present disclosure provides a data processing method and device, electronic equipment and storage medium, and relates to the technical field of data processing. The method comprises the following steps: obtaining a first parameter generated by a parameter adjustment module in a target state; the first parameter is used for data processing performed by a data processing module; determining a first evaluation result of the first parameter; the first evaluation result is used for evaluating the effect of data processing under the first parameter; generating a second parameter according to the first evaluation result and the first parameter; and updating the target state according to the first evaluation result and the second parameter. The second parameter is an additional generated parameter. According to the first evaluation result and the second parameter, the target state is updated, the characteristics used for updating the target state are enriched, the updating effect of the target state is improved, the parameter adjustment module generates better parameters, the data processing module uses better parameters for data processing, and the data processing effect of the data processing module is indirectly improved.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and in particular to a data processing method, apparatus, electronic device and storage medium. Background Technology

[0002] The data processing module can process data. It includes adjustable parameters. By adjusting these adjustable parameters through the parameter adjustment module, the data processing performance of the data processing module can be improved. Summary of the Invention

[0003] This disclosure provides a data processing method, apparatus, electronic device, and storage medium to solve problems in the related art.

[0004] A first aspect of this disclosure provides a data processing method, the method comprising: The first parameter generated by the parameter adjustment module in the target state is obtained; the first parameter is used by the data processing module to perform data processing. Determine the first evaluation result for the first parameter; the first evaluation result is used to evaluate the effectiveness of data processing under the first parameter. The second parameter is generated based on the first evaluation result and the first parameter; The target state is updated based on the first evaluation result and the second parameter.

[0005] The second parameter is an additional parameter generated. Based on the first evaluation result and the second parameter, the target state is updated, which enriches the features used to update the target state, thereby improving the update effect of the target state. This helps the parameter adjustment module generate better parameters, enabling the data processing module to use better parameters for data processing, and indirectly improving the data processing effect of the data processing module.

[0006] In this disclosure, the first evaluation result for determining the first parameter includes: Replace the original parameters in the data processing module with the first parameter. The original parameters are the parameters generated by the parameter adjustment module in the state before the target state. Based on the first parameter, the data processing module performs data processing on the first data to obtain the second data; Based on the second set of data, the first evaluation result is determined.

[0007] The second data is obtained by processing the first data based on the first parameter. The second data can reflect the effect of the data processing on the first data by the first parameter, and thus the first evaluation result of the first parameter can be indirectly determined through the second data. Determining the first evaluation result of the first parameter indirectly through the second data eliminates the need to evaluate the first parameter itself.

[0008] In this disclosure, the determination of the first evaluation result based on the second data includes: Get the labels of the first data; The first evaluation result is determined based on the difference between the second data and the label.

[0009] The difference reflects the processing effect of the data processing module on the first data. Since the data processing module processes the first data based on the first parameter, the difference can reflect the processing effect of the first parameter on the first data. The difference analysis is relatively simple. Determining the first evaluation result through the difference can reduce the complexity of determining the first evaluation result, thereby improving the efficiency of determining the first evaluation result.

[0010] In this disclosure, the generation of the second parameter based on the first evaluation result and the first parameter includes: Evaluation indicators for obtaining the first evaluation result; According to the evaluation indicators, a set of differences is determined. The set of differences includes at least one difference. Each difference corresponds to one of the evaluation indicators. The difference is the difference between the label of the first data and the second data. The second data is obtained by the data processing module performing data processing on the first data under the first parameters. The second parameter is generated based on the difference set and the first parameter.

[0011] Variance analysis can be conducted based on evaluation indicators, with different indicators corresponding to different variances. Using evaluation indicators to conduct variance analysis can guide the analysis and improve the accuracy of variance identification.

[0012] In this disclosure, the set of differences is determined according to the evaluation indicators, including: In response to a difference set including one difference, a virtual evaluation index is constructed based on the target state; Determine the virtual differences between the second data and the labels under the virtual evaluation index; The difference set is determined based on the evaluation indicators and the virtual evaluation indicators; the difference set includes the actual difference and the virtual difference.

[0013] When there is only one evaluation indicator, there is also only one difference. Due to the small number of differences, the dimension of the difference set is small. By constructing a virtual evaluation indicator, the dimension of the evaluation indicator can be expanded, and the dimension of the difference set can be increased, thereby providing more features for the generation of the second parameter and improving the accuracy of the second parameter generation.

[0014] In this disclosure, generating the second parameter based on the difference set and the first parameter includes: The first parameter is normalized to obtain the normalized parameter; A correlation analysis was performed on the normalized parameters and the difference set to obtain the correlation matrix between the normalized parameters and the difference set; The second parameter is generated based on the correlation matrix.

[0015] Normalization can reduce the data dimensionality of the first parameter, thereby reducing the complexity of correlation analysis, improving the efficiency of correlation analysis, and thus improving the generation efficiency of the second parameter.

[0016] In this disclosure, the correlation matrix includes multiple correlation sub-matrices, and the second parameter includes a third parameter and a fourth parameter. The second parameter is generated based on the correlation matrix, including: Identify the target correlation submatrix with the highest correlation among multiple correlation submatrices; Parametric analysis is performed on the correlations in the target correlation submatrix to obtain the third parameter; A fourth parameter is obtained by performing parameter analysis on the correlation within a preset range; the preset range is the range based on the correlation matrix.

[0017] The generation methods for the third and fourth parameters enable the generation of the second parameter in multiple ways, thus expanding the generation methods of the second parameter and increasing its richness.

[0018] In this disclosure, updating the target state based on the first evaluation result and the second parameter includes: The second evaluation result is determined by the second parameter; The target status is updated based on the results of the first and second evaluations.

[0019] The second evaluation result provides additional features for updating the target state, thereby improving the update effect of the target state.

[0020] A second aspect of this disclosure provides a data processing apparatus, the apparatus comprising: The acquisition unit is used to acquire the first parameter generated by the parameter adjustment module in the target state; the first parameter is used by the data processing module to perform data processing. A determining unit is used to determine a first evaluation result for a first parameter; the first evaluation result is used to evaluate the effect of data processing under the first parameter. The generation unit is used to generate the second parameter based on the first evaluation result and the first parameter; The update unit is used to update the target state based on the first evaluation result and the second parameter.

[0021] In this disclosure, the determining unit is also used for: Replace the original parameters in the data processing module with the first parameter. The original parameters are the parameters generated by the parameter adjustment module in the state before the target state. Based on the first parameter, the data processing module performs data processing on the first data to obtain the second data; Based on the second set of data, the first evaluation result is determined.

[0022] In this disclosure, the determining unit is also used for: Get the labels of the first data; The first evaluation result is determined based on the difference between the second data and the label.

[0023] In this disclosure, the generating unit is also used for: Evaluation indicators for obtaining the first evaluation result; According to the evaluation indicators, a set of differences is determined. The set of differences includes at least one difference. Each difference corresponds to one of the evaluation indicators. The difference is the difference between the label of the first data and the second data. The second data is obtained by the data processing module performing data processing on the first data under the first parameters. The second parameter is generated based on the difference set and the first parameter.

[0024] In this disclosure, the generating unit is also used for: In response to a difference set including one difference, a virtual evaluation index is constructed based on the target state; Determine the virtual differences between the second data and the labels under the virtual evaluation index; The difference set is determined based on the evaluation indicators and the virtual evaluation indicators; the difference set includes the actual difference and the virtual difference.

[0025] In this disclosure, the generating unit is also used for: The first parameter is normalized to obtain the normalized parameter; A correlation analysis was performed on the normalized parameters and the difference set to obtain the correlation matrix between the normalized parameters and the difference set; The second parameter is generated based on the correlation matrix.

[0026] In this disclosure, the correlation matrix includes multiple correlation sub-matrices, the second parameter includes a third parameter and a fourth parameter, and the generating unit is further used for: Identify the target correlation submatrix with the highest correlation among multiple correlation submatrices; Parametric analysis is performed on the correlations in the target correlation submatrix to obtain the third parameter; A fourth parameter is obtained by performing parameter analysis on the correlation within a preset range; the preset range is the range based on the correlation matrix.

[0027] In this disclosure, the updating unit is also used for: The second evaluation result is determined by the second parameter; The target status is updated based on the results of the first and second evaluations.

[0028] A third aspect of this disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the methods described in the first aspect of this disclosure.

[0029] A fourth aspect of this disclosure provides a non-transitory computer-readable storage medium that, when instructions in the storage medium are executed by a processor of a mobile terminal, enables the mobile terminal to perform the methods described in the first aspect of this disclosure.

[0030] According to a fifth aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the methods described in the embodiments of the first aspect of this disclosure.

[0031] A sixth aspect of this disclosure provides a chip including one or more interfaces and one or more processors; the interfaces are configured to receive signals from the memory of an electronic device and send signals to the processors, the signals including computer instructions stored in the memory, which, when executed by the processors, cause the electronic device to perform the methods described in the first aspect of this disclosure.

[0032] In summary, the data processing method proposed in this disclosure includes obtaining a first parameter generated by a parameter adjustment module under a target state; the first parameter being used by a data processing module to perform data processing; determining a first evaluation result of the first parameter; the first evaluation result being used to evaluate the effect of data processing under the first parameter; generating a second parameter based on the first evaluation result and the first parameter; and updating the target state based on the first evaluation result and the second parameter. The second parameter is an additionally generated parameter. Updating the target state based on the first evaluation result and the second parameter enriches the features used to update the target state, thereby improving the update effect of the target state. This helps the parameter adjustment module generate better parameters, enabling the data processing module to utilize better parameters for data processing, indirectly improving the data processing effect of the data processing module.

[0033] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0034] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0035] Figure 1 A flowchart of a data processing method provided in an embodiment of this disclosure; Figure 2 A flowchart illustrating a method for determining an evaluation result provided in an embodiment of this disclosure; Figure 3 A flowchart illustrating a state update method provided in an embodiment of this disclosure; Figure 4 A flowchart illustrating a method for constructing a residual provided in an embodiment of this disclosure; Figure 5 A flowchart illustrating a method for generating development candidates provided in an embodiment of this disclosure; Figure 6 A flowchart illustrating a method for generating candidate explorations, as provided in this embodiment of the disclosure; Figure 7 This is a schematic diagram of the structure of a data processing apparatus provided in an embodiment of the present disclosure; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure; Figure 9 This is a schematic diagram of the structure of a chip provided in an embodiment of the present disclosure. Detailed Implementation

[0036] Some embodiments of this disclosure will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. Various changes, modifications, and equivalents of the methods, apparatus, and / or systems described herein will become apparent upon understanding this disclosure. For example, the order of operations described herein is merely illustrative and is not limited to those orders set forth herein, but can be changed as will become apparent upon understanding this disclosure, except for operations that must be performed in a particular order. Furthermore, for clarity and brevity, descriptions of features known in the art may be omitted.

[0037] The embodiments described in the following examples of this disclosure are not representative of all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0038] The data processing module can process data. The data processing module includes adjustable parameters. By adjusting the adjustable parameters in the data processing module through the parameter adjustment module, the performance of the data processing module in processing data can be adjusted.

[0039] Therefore, to address the problems existing in related technologies, this disclosure proposes a data processing method. This method includes obtaining a first parameter generated by a parameter adjustment module under a target state; the first parameter being used by a data processing module to perform data processing; determining a first evaluation result of the first parameter; the first evaluation result being used to evaluate the effect of data processing under the first parameter; generating a second parameter based on the first evaluation result and the first parameter; and updating the target state based on the first evaluation result and the second parameter. The second parameter is an additionally generated parameter. Updating the target state based on the first evaluation result and the second parameter enriches the features used to update the target state, thereby improving the update effect of the target state. This helps the parameter adjustment module generate better parameters, enabling the data processing module to utilize better parameters for data processing, indirectly improving the data processing effect of the data processing module.

[0040] This disclosure is not exhaustive, but merely illustrative of some embodiments, and is not intended to limit the scope of protection of this disclosure. Unless otherwise specified, each step in a particular embodiment can be implemented as an independent embodiment, and the steps can be arbitrarily combined. For example, a solution after removing some steps in a particular embodiment can also be implemented as an independent embodiment, and the order of the steps in a particular embodiment can be arbitrarily interchanged. Furthermore, the optional implementation methods in a particular embodiment can be arbitrarily combined; moreover, the embodiments can be arbitrarily combined, for example, some or all steps of different embodiments can be arbitrarily combined, and a particular embodiment can be arbitrarily combined with the optional implementation methods of other embodiments.

[0041] In each of the disclosed embodiments, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions of the embodiments are consistent and can be referenced by each other. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0042] The terminology used in the embodiments of this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the scope of this disclosure.

[0043] In this embodiment of the disclosure, unless otherwise stated, elements expressed in the singular form, such as "a," "an," "the," "the aforementioned," "the," "this," etc., can mean "one and only one," or "one or more," "at least one," etc. For example, when using articles such as "a," "an," "the," etc. in translation, the noun following the article can be understood as either a singular or a plural expression.

[0044] In this disclosure, the terms “in response to…”, “in response to determining…”, “in the case of…”, “when…”, “if…”, “if…”, etc., can be used interchangeably.

[0045] In this disclosure, terms such as “greater than”, “greater than or equal to”, “not less than”, “more than”, “more than or equal to”, “not less than”, “higher than”, “higher than or equal to”, “not lower than”, and “above” can be used interchangeably, as can terms such as “less than”, “less than or equal to”, “not greater than”, “less than”, “less than or equal to”, “not more than”, “lower than”, “lower than or equal to”, “not higher than”, and “below”.

[0046] The prefixes such as "first" and "second" in the embodiments of this disclosure are only for distinguishing different descriptive objects and do not constitute restrictions on the position, order, priority, number or content of the descriptive objects. For the description of the descriptive objects, please refer to the description in the claims or the context of the embodiments. The use of prefixes should not constitute unnecessary restrictions.

[0047] In the embodiments of this disclosure, "multiple" refers to two or more.

[0048] In the embodiments disclosed herein, terms such as “import”, “input”, and “read in” can be used interchangeably.

[0049] In this disclosure, the terms "device" and "apparatus" can be interpreted as physical or virtual, and their names are not limited to those described in the embodiments. Terms such as "apparatus," "equipment," "device," "circuit," "network element," "node," "function," "unit," "section," "system," "network," "chip," "chip system," "entity," and "subject" can be used interchangeably.

[0050] In this disclosure, the collection, storage, use, processing, transmission, provision, and application of user personal information comply with relevant laws and regulations and do not violate public order and good morals. It should be noted that personal information from users should be collected for legitimate and reasonable purposes and not shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should be conducted only after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0051] Figure 1 This is a flowchart illustrating a data processing method provided in an embodiment of this disclosure. This method can be applied to application scenarios such as smart terminals, for example, executed by a terminal integrating state update and data processing functions or a processor within a terminal, or executed by other devices suitable for performing state update and data processing functions; this disclosure does not limit its application. Figure 1 As shown, the data processing method includes steps S101-S104.

[0052] Step S101: Obtain the first parameter generated by the parameter adjustment module in the target state; the first parameter is used by the data processing module to perform data processing.

[0053] In embodiments of this disclosure, the first parameter can be a parameter deployed in the data processing module. The data processing module can use the first parameter to process data. The parameter adjustment module is used to generate the parameters used by the data processing module, and the state of the parameter adjustment module corresponds one-to-one with the parameters. Embodiments of this disclosure do not limit the type of the first parameter. Embodiments of this disclosure do not limit the data processing function of the data processing module, nor do they limit the data types processed by the data processing module.

[0054] The first parameter can be deployed in the data processing module, enabling the data processing module to process the data based on the first parameter, thereby improving the data processing effect of the data processing module.

[0055] Step S102: Determine the first evaluation result of the first parameter; the first evaluation result is used to evaluate the effect of data processing under the first parameter.

[0056] In the embodiments of this disclosure, the data processing module processes the data using a first parameter to obtain a data processing result. This data processing result can be evaluated to obtain an evaluation result. Since the data processing result is based on the first parameter, the evaluation result can reflect the first evaluation result of the first parameter. Therefore, the evaluation result of the data processing result can be determined as the first evaluation result.

[0057] In the embodiments of this disclosure, the evaluation type and evaluation criteria for the first evaluation result are not limited.

[0058] The first evaluation result reflects the effect of data processing under the first parameter. The first evaluation result can determine the update direction of the target state, so that the parameter adjustment module generates parameters based on the target state tending towards a better data processing effect, thereby improving the accuracy of the parameters generated by the parameter adjustment module.

[0059] Step S103: Generate the second parameter based on the first evaluation result and the first parameter.

[0060] In the embodiments of this disclosure, the specific method of generating the second parameter is not limited. The second parameter is a parameter additionally generated based on the first evaluation result and the first parameter. The second parameter can be used to update the target state, and provides a data basis for updating the target state.

[0061] Step S104: Update the target state based on the first evaluation result and the second parameter.

[0062] In the embodiments of this disclosure, the specific method of updating the target state is not limited. The second parameter can provide additional features for updating the target state. By updating the target state according to the first evaluation result and the second parameter, the features used to update the target state are enriched, thereby improving the update effect of the target state. This helps the parameter adjustment module generate better parameters, enabling the data processing module to use better parameters for data processing, and indirectly improving the data processing effect of the data processing module.

[0063] According to the data processing method proposed in this disclosure, the method includes obtaining a first parameter generated by a parameter adjustment module under a target state; the first parameter being used by a data processing module to perform data processing; determining a first evaluation result of the first parameter; the first evaluation result being used to evaluate the effect of data processing under the first parameter; generating a second parameter based on the first evaluation result and the first parameter; and updating the target state based on the first evaluation result and the second parameter. The second parameter is an additionally generated parameter. Updating the target state based on the first evaluation result and the second parameter enriches the features used to update the target state, thereby improving the update effect of the target state. This helps the parameter adjustment module generate better parameters, enabling the data processing module to use better parameters for data processing, indirectly improving the data processing effect of the data processing module.

[0064] Figure 2 This is a flowchart illustrating a method for determining evaluation results as proposed in this disclosure. Based on Figure 1 The embodiment shown, Figure 2 This may include the following steps: Step S201: Replace the original parameters in the data processing module with the first parameters. The original parameters are the parameters generated by the parameter adjustment module in the state before the target state.

[0065] In embodiments of this disclosure, the state preceding the target state can be the previous state. If the target state is an initial state, the original parameters can be the initial parameters of the data processing module.

[0066] By replacing the original parameters with the first parameter, dynamic updates to the data processing module are achieved, thereby improving the data processing efficiency of the data processing module.

[0067] Step S202: Based on the first parameter, the data processing module performs data processing on the first data to obtain the second data.

[0068] In the embodiments of this disclosure, the data types of the first data and the second data are not limited, nor is the process by which the data processing module processes the first data using the first parameter limited. The first parameter may contain the execution logic for data processing, and processing the first data using the first parameter can improve the data processing effect.

[0069] Step S203: Determine the first evaluation result based on the second data.

[0070] In the embodiments of this disclosure, the second data is obtained by processing the first data based on the first parameter. The second data can reflect the effect of the data processing of the first data by the first parameter, thereby indirectly determining the first evaluation result of the first parameter through the second data. Determining the first evaluation result of the first parameter indirectly through the second data eliminates the need to evaluate the first parameter itself.

[0071] For example, when determining the first evaluation result based on the second data, the following methods may be used, but are not limited to: obtaining the labels of the first data; and determining the first evaluation result based on the difference between the second data and the labels.

[0072] In the embodiments of this disclosure, the label can be reference data obtained by processing the first data, and the label reflects the correct data obtained by processing the first data; the embodiments of this disclosure do not limit the type of label. The embodiments of this disclosure do not limit the specific method of difference analysis between the second data and the label. The first evaluation result can also be determined by the similarity between the second data and the label. The difference can be determined as the first evaluation result.

[0073] The difference reflects the processing effect of the data processing module on the first data. Since the data processing module processes the first data based on the first parameter, the difference can reflect the processing effect of the first parameter on the first data. The difference analysis is relatively simple. Determining the first evaluation result through the difference can reduce the complexity of determining the first evaluation result, thereby improving the efficiency of determining the first evaluation result.

[0074] For example, when generating the second parameter based on the first evaluation result and the first parameter, the following methods can be used, but are not limited to: obtaining the evaluation index of the first evaluation result; determining the difference set according to the evaluation index, the difference set including at least one difference, the difference corresponding one-to-one with the evaluation index, the difference being the difference between the label of the first data and the second data, the second data being obtained by the data processing module performing data processing on the first data under the first parameter; and generating the second parameter based on the difference set and the first parameter.

[0075] In the embodiments of this disclosure, the type and number of evaluation indicators are not limited. Difference analysis can be performed based on evaluation indicators, with different indicators corresponding to different differences. Performing difference analysis according to evaluation indicators can indicate the direction of the analysis and improve the accuracy of difference determination.

[0076] For example, when determining the set of differences according to the evaluation index, it can be implemented in the following ways, but not limited to: in response to the inclusion of a difference in the set of differences, constructing a virtual evaluation index based on the target state; determining the virtual difference between the second data and the label under the virtual evaluation index; determining the set of differences according to the evaluation index and the virtual evaluation index; the set of differences includes the differences and the virtual differences.

[0077] In the embodiments of this disclosure, the specific construction method of the virtual evaluation index and the specific determination method of the virtual difference are not limited. When there is only one evaluation index, there is also only one difference. Since the number of differences is small, the dimension of the difference set is small. By constructing a virtual evaluation index, the dimension of the evaluation index can be expanded, and the dimension of the difference set can be increased, thereby providing more features for the generation of the second parameter and improving the accuracy of the second parameter generation.

[0078] For example, when generating the second parameter based on the difference set and the first parameter, the following methods can be used, but are not limited to: normalizing the first parameter to obtain a normalized parameter; performing a correlation analysis on the normalized parameter and the difference set to obtain a correlation matrix between the normalized parameter and the difference set; and generating the second parameter based on the correlation matrix.

[0079] In the embodiments of this disclosure, the specific method of normalization processing is not limited, nor is the specific method of correlation analysis limited. The specific method of generating the second parameter is not limited.

[0080] Normalization can reduce the data dimensionality of the first parameter, thereby reducing the complexity of correlation analysis, improving the efficiency of correlation analysis, and thus improving the generation efficiency of the second parameter.

[0081] For example, the correlation matrix includes multiple correlation sub-matrices, and the second parameter includes a third parameter and a fourth parameter. When generating the second parameter based on the correlation matrix, it can be implemented in the following ways, but is not limited to: determining the target correlation sub-matrix with the highest correlation among the multiple correlation sub-matrices; performing parameter analysis on the correlation in the target correlation sub-matrix to obtain the third parameter; performing parameter analysis on the correlation within a preset range to obtain the fourth parameter; the preset range is the range based on the correlation matrix.

[0082] In the embodiments of this disclosure, the method of dividing the correlation submatrix is ​​not limited. The number of correlation submatrices included in the correlation matrix is ​​not limited. The specific analysis method for parameter analysis is not limited.

[0083] The generation methods for the third and fourth parameters enable the generation of the second parameter in multiple ways, thus expanding the generation methods of the second parameter and increasing its richness.

[0084] For example, when updating the target state based on the first evaluation result and the second parameter, the following methods can be used, but are not limited to: determining the second evaluation result of the second parameter; and updating the target state based on the first evaluation result and the second evaluation result.

[0085] In the embodiments of this disclosure, the method for determining the second evaluation result is the same as the method for determining the first evaluation result. For details on the method for determining the second evaluation result, please refer to the method for determining the first evaluation result in the above embodiments. The method for determining the second evaluation result will not be described in detail in the embodiments of this disclosure.

[0086] The second evaluation result provides additional features for updating the target state, thereby improving the update effect of the target state.

[0087] In the embodiments of this disclosure, the application scenarios of these embodiments may include, but are not limited to, black-box optimization scenarios, evolutionary computation scenarios, derivative-free optimization scenarios, swarm intelligence optimization scenarios, and multi-objective optimization scenarios. Specifically, scalar black-box optimization may include engineering parameter tuning, simulation-driven design, automatic machine learning hyperparameter search, algorithm configuration, and high-dimensional continuous optimization; structured objective optimization may include, for example, path planning, robot trajectory planning, control parameter optimization, and engineering design with decomposable constraint costs; multi-objective optimization may include frontier search with low and high objective numbers, hypervolume optimization, and multi-index tradeoffs under complex constraints; candidate generation modules combining optimizers; image algorithm debugging requiring cross-scene optimization, such as image quality tuning, automatic white balance tuning, automatic exposure, noise reduction, sharpening, and color correction in image signal processors (ISPs); and optimization systems requiring limited evaluation budgets and where gradients cannot be reliably obtained.

[0088] In the embodiments of this disclosure, the embodiments can be further extended to high-dimensional continuous function optimization scenarios, such as test functions, engineering simulation functions, and algorithm parameter searches, to solve the problem that the evaluation function only returns scalar values ​​and cannot provide analytical gradients or component losses. The embodiments of this disclosure can be further extended to path planning scenarios, where the path is represented by high-dimensional parameters, and the objective is jointly determined by structured costs such as flight length, terrain risk, smoothness, and collision penalty, solving the problem that blind covariance learning cannot fully utilize local structures within a short budget. The embodiments of this disclosure can be further extended to multi-objective optimization scenarios, where the evaluation function returns multiple objective values, and the objective vector can be used as a residual. The embodiments of this disclosure can be further extended to industrial ISP image quality joint parameter tuning scenarios, where candidate points can represent high-dimensional ISP configurations, and evaluation components can include image quality indicators such as noise, brightness, color, white balance, exposure, contrast, sharpening, texture, dynamic range, artifacts, and local overexposure / underexposure, and can also include scene-level errors under different scenes, lighting conditions, ISOs, skin tones, or shooting modes. This disclosure can be further extended to scenarios with limited evaluation budgets. When the cost of each simulation, experiment, model training, or real system test is high, this disclosure can generate additional candidates using existing recent samples to improve the return per unit evaluation budget.

[0089] In embodiments of this disclosure, without acquiring analytical gradients, training the policy network offline, or replacing the master optimizer, residual Jacobian is estimated online using recent evaluation trajectories, and optimizations for local alignment candidates are generated based on the Jacobian.

[0090] In the embodiments of this disclosure, a black-box optimization scenario is used as an example. In the embodiments of this disclosure, the parameter adjustment module can be an optimizer, the data processing module can be a black box, the first parameter can be a native candidate, the second parameter can be an additionally generated candidate, the correlation matrix can be a residual Jacobian matrix, the third parameter can be a development candidate, and the fourth parameter can be an exploration candidate.

[0091] For example, in a black-box optimization system with a candidate generation-evaluation-update process, the system generates native candidates in each generation by the main optimizer, calls a black-box function or simulator to obtain evaluation results, and updates the internal state by the main optimizer according to selection, survival, or ranking rules. In this process, the scheme of this embodiment reads the recent evaluation window, fits the local residual Jacobian, generates a small number of additional candidates, evaluates these candidates together with the main optimizer's native candidates, and then returns all evaluation results to the main optimizer for processing.

[0092] In embodiments of this disclosure, based on Figure 1 The illustrated embodiments, such as Figure 3 As shown, Figure 3The flowchart illustrates the state update method provided in this embodiment of the disclosure. This embodiment may include primary optimizer sampling, black-box evaluation, recent window maintenance, residual construction, Jacobi fitting, Gauss-Newton (GN) candidate generation, CMIX candidate generation, and primary optimizer update.

[0093] In embodiments of this disclosure, the generation of native candidates by the primary optimizer may include, in the g-th generation, the primary optimizer A generating a batch of native candidates based on its current state. For example, the primary optimizer may sample according to mean, step size, and covariance; the primary optimizer may generate candidates according to difference mutation; and a multi-objective primary optimizer may generate candidates according to its crossover mutation rules. Evaluating and maintaining the recent window may include performing black-box evaluation on the candidates and saving the most recent n evaluated samples to form an evaluation window. If the evaluation function provides objective components or multi-objective vectors, the corresponding components are saved synchronously. Primary optimizer normalization may include converting candidate points into normalized values ​​according to the current local scale of the primary optimizer, making the scales of different dimensions and different primary optimizers comparable. Constructing residuals may include constructing residuals according to the type of evaluation information: if there are components or multi-objective vectors, the normalized components, square root components, or standardized objectives are used directly; if only scalar values ​​are available, a linear or quadratic local model is fitted, and pseudo-residuals are generated from the slope or curvature direction of the model.

[0094] In embodiments of this disclosure, online fitting of the residual Jacobian may include fitting a linear relationship "normalized to residuals" using local regression over a recent window to obtain the residual Jacobian J, and then deriving the local metric G from J. Bounded development candidates are generated. Using the current best point within the window and its residual as a benchmark, a damping step size GN is calculated to obtain candidate points. To avoid disrupting the master optimizer step size control, the candidate is pruned during master optimizer whitening to ensure its step size does not exceed a preset radius. Generating inverse curvature CMIX exploration candidates may include converting the same local metric into a regularized inverse curvature covariance and performing trace-preserving mixing with the master optimizer covariance or candidate sampling covariance. This operation keeps the total sampling energy constant, only changing the orientation shape, so that locally flat directions obtain more variance and sharp directions obtain less variance.

[0095] In embodiments of this disclosure, candidate evaluation and main optimizer updates may include evaluating GN candidates and CMIX candidates, merging their evaluation results with the main optimizer's native candidates, and then having the main optimizer perform the original selection, survival, covariance adaptation, step size control, restart, and boundary handling. Embodiments of this disclosure do not alter the aforementioned main optimizer logic. Iterative execution until the budget is exhausted may include refitting J using the latest recent window in each generation, discarding old samples, thus ensuring that the local model always serves only the current search region.

[0096] In embodiments of this disclosure, a residual least squares perspective is used. When the target consists of multiple non-negative components, the residual vector may include: .

[0097] in, For the first The transformed residual components Let be the parameter vector to be optimized. For the first Each residual component This is the transpose of the matrix.

[0098] In embodiments of this disclosure, optimization objectives may include:

[0099] in, Let be the black-box objective loss function.

[0100] In embodiments of this disclosure, the residual Jacobian matrix may include:

[0101] in, Let be the residual Jacobian matrix. for OK, The space of real matrices in columns, for A set of.

[0102] In embodiments of this disclosure, local Jacobian estimation is performed in a black-box scenario. Within the recent window, writing candidate points into the main optimizer for normalization may include:

[0103] in, For the first in the original parameter space One sample point, For the whitening coordinate space, the first A sample vector, It is the inverse matrix. The mean vector of the distribution maintained by the master optimizer.

[0104] In embodiments of this disclosure, the local regression fitting residuals may include:

[0105] in, This is the bias vector.

[0106] In embodiments of this disclosure, local metrics may include:

[0107] in, For local metric matrices, Let be the residual Jacobian matrix. It is the transpose of the residual Jacobian matrix.

[0108] In embodiments of this disclosure, developing a matrix to utilize candidates may include setting the optimal point within the window as... The corresponding residual is The damping step size can include:

[0109] in, To update the step size vector, The damping coefficient is... It is a unit square matrix.

[0110] In embodiments of this disclosure, injection candidates may include:

[0111] in, As a candidate point for damping, As the search center point, This is the step scaling factor.

[0112] In embodiments of this disclosure, to ensure that the candidate scale is consistent with the native sampling of the main optimizer, bounded pruning during whitening may include:

[0113] in, This is the damping candidate offset vector in the whitening coordinate system. This is the preset radius scaling factor. The total dimension of the variable to be optimized.

[0114] In embodiments of this disclosure, CMIX inverse curvature exploration candidates are obtained by converting the same local metric into an inverse curvature matrix, which may include:

[0115] in, It is the inverse curvature matrix. This is the mixed damping coefficient.

[0116] In embodiments of this disclosure, trace-preserving mixing of the primary optimizer covariance may include:

[0117] in, To preserve the new sampling covariance matrix obtained after trace mixing, For mixed weighting coefficients, The native covariance matrix of the primary optimizer. This is the trace operation of a matrix.

[0118] In the embodiments of this disclosure,

[0119] In the embodiments of this disclosure, the total sampling energy is kept constant, and only the variance in different directions is redistributed.

[0120] In embodiments of this disclosure, the dimension-adaptive weights and hybrid weights can be set as follows:

[0121] in, For the residual dimension, This represents the number of samples in the recent window. The search space dimension is defined by `clip`, which is the truncation function. Based on the weight constant, The minimum allowed value for mixed weights. This represents the maximum allowed value for the mixed weights.

[0122] In embodiments of this disclosure, based on Figure 1 The illustrated embodiments, such as Figure 4 As shown, Figure 4 The flowchart illustrates a residual construction method provided in this embodiment of the present disclosure. When there are structured components or multi-objective vectors, standardization or square root transformation is performed directly; when there are only scalar objectives, a linear or quadratic local model is used to construct pseudo-residuals.

[0123] In embodiments of this disclosure, the observation component model may include using the square root component, standardized component, or shifted component value as the residual when the objective function exposes a non-negative scene loss, constraint loss, cost component, or multi-objective vector. This model is suitable for path planning, multi-objective optimization, and decomposable engineering objectives. The linear scalar model may include fitting a locally linear model over a recent window and using the normalized linear prediction value as the one-dimensional residual when the evaluation function only returns a scalar f(x). This model provides robust local slope estimation and is suitable for generating conservative GN candidates.

[0124] In embodiments of this disclosure, the quadratic scalar mode may include fitting a quadratic model over a recent window, performing eigenvalue decomposition on the quadratic term matrix, and constructing curvature alignment residuals based on eigenvalues ​​and eigenvectors when multiple local curvature directions need to be extracted from scalar values. This mode provides multiple curvature alignment directions for CMIX exploration.

[0125] In embodiments of this disclosure, based on Figure 1 The illustrated embodiments, such as Figure 5 As shown, Figure 5This is a flowchart illustrating a candidate generation method provided in an embodiment of this disclosure. The GN candidate generation method may include selecting the current best point within a recent window, calculating the damped Gauss-Newton direction based on the residual Jacobian, and pruning the candidate step size during the main optimizer whitening process to avoid disrupting the original step size control of the main optimizer. The GN candidate follows a descent direction on the fitted residual model. When λ is non-zero and λ>0, the decrease in the local model is non-positive, and therefore it can be used as a development candidate. To ensure safe injection, the candidate is pruned in the whitening coordinates of the main optimizer so that its scale is consistent with the normal sampling scale of the main optimizer, thus avoiding a local model candidate from dominating the entire optimization process.

[0126] In embodiments of this disclosure, based on Figure 1 The illustrated embodiments, such as Figure 6 As shown, Figure 6 This flowchart illustrates a method for generating exploration candidates according to an embodiment of this disclosure. The CMIX exploration candidate generation method may include converting local metrics into inverse curvature matrices and performing trace-preserving mixing with the main optimizer covariance to keep the total sampling energy constant, adjusting only the variance allocation in different directions. CMIX does not increase or decrease the overall sampling energy; it only redistributes directional variance based on local residual curvature. The residual model assumes that directions with flatter changes or insufficient exploration receive larger variances, while directions with sharper changes receive smaller variances. This mechanism uses the same J as GN candidates, thus ensuring a consistent geometric origin for both exploration and development candidates.

[0127] In embodiments of this disclosure, the primary optimizer includes candidate generation, evaluation result reception, and state update processes. The tasks of the primary optimizer may include native candidate generation, sample selection or multi-objective survival, covariance, mutation scale, step size, or population state updates, boundary handling and restart, and global diversity maintenance. Embodiments of this disclosure may provide a small number of additional candidate or temporary candidate sampling metrics, which can be used as an external module of the population optimizer.

[0128] In embodiments of this disclosure, w=0 can be set to use only GN bounded candidates and CMIX can be omitted. Embodiments of this disclosure are applicable to scenarios where scalar targets, local curvature spectrum noise is high, or global basin discovery remains a major bottleneck.

[0129] In embodiments of this disclosure, GN candidates and CMIX candidates can be used, which is suitable for scenarios where the recent window can form stable local residual geometry, such as structured objectives and multi-objective optimization.

[0130] In embodiments of this disclosure, GN and CMIX candidates can be used as additional candidates for DE in the evaluation, and the evaluation results are still processed by the selection rules of DE.

[0131] In embodiments of this disclosure, the target vector can be directly used as the residual, and the GN and CMIX candidates are evaluated and then processed by the survival mechanism of the target optimization algorithm.

[0132] In the embodiments of this disclosure, the embodiments of this disclosure can solve high-dimensional problems such as path planning, and use path quality components, constraint penalties or target decomposition items as residuals to improve local improvement capabilities under short budgets.

[0133] In embodiments of this disclosure, the residual-direction relationship in the recent evaluation window can be transformed into a Jacobian matrix J, so that local geometry that was not originally utilized by the main optimizer can be used for candidate generation.

[0134] In embodiments of this disclosure, the same J can be used simultaneously for both development candidate generation and exploration candidate generation, reducing module complexity and providing a unified geometric source for development and exploration.

[0135] In embodiments of this disclosure, the main optimizer may not be replaced, and the selection, survival, covariance adaptation, or step size control may not be changed; GN candidates are bounded pruning, and CMIX are trace-preserving mixing, reducing the risk of disrupting the main optimizer at the interface level. Embodiments of this disclosure can construct residuals and fit the Jacobian online.

[0136] In embodiments of this disclosure, local geometry is estimated only within the current optimization instance, without the need to train distributions, policy networks, or cross-task parameter transfers.

[0137] In the embodiments of this disclosure, in the ISP image quality parameter tuning, indicators such as noise, brightness, color, white balance, exposure, sharpening, and dynamic range can all be included as residual components in the same optimization framework.

[0138] In embodiments of this disclosure, by disabling GN or CMIX, or setting w=0, it can be directly verified that the gains come from additional candidate generation rather than changes in the master optimizer protocol, population size, or evaluation budget.

[0139] In embodiments of this disclosure, candidate points can be configured as high-dimensional ISPs, and noise, brightness, color, white balance, exposure, sharpening, texture, dynamic range, artifacts, and image quality errors under different scenes together constitute the residual vector. The system uses this residual vector to estimate J and simultaneously improves candidate injection and sampling exploration.

[0140] In the embodiments of this disclosure, in the scalar black-box optimization function, the objective function only returns f(x). The main optimizer can be CMA-ES. Each generation collects the most recent n evaluation samples, and normalizes the samples using the current mean and scale of the main optimizer. If a linear scalar mode is used, a locally linear model is fitted and a one-dimensional residual is constructed; if a quadratic scalar mode is used, a quadratic model is fitted and a multi-dimensional residual is constructed from the feature directions. J is then obtained through ridge regression, GN candidates are calculated, and bounded injection is performed; when the curvature spectrum is stable, CMIX candidates are further used. After evaluation, all samples are processed by the original selection and covariance update of CMA-ES.

[0141] In embodiments of this disclosure, during path planning, candidate x represents high-dimensional path parameters, such as 1135-dimensional spherical path encoding. An evaluation function returns structured components such as terrain risk, collision penalty, smoothness, or energy consumption. The scheme normalizes these components as residuals, fits J online, and generates GN and CMIX candidates. Because the target components of path planning have explicit physical meaning, the residuals are generally more stable than scalar pseudo-residuals, thus CMIX can be enabled earlier to more effectively distribute search variance within locally feasible corridors.

[0142] In embodiments of this disclosure, in multi-objective tasks, the objective vector can be directly used as the residual. The main optimizer can be a multi-objective optimization algorithm. Each generation generates native candidates by the main optimizer. The optimizer can fit J based on the recent objective vector to generate a small number of GN or CMIX candidates. After evaluation, the original multi-objective survival operator decides whether to retain them. Embodiments of this disclosure do not change the multi-objective optimization ranking, hypervolume contribution, or adaptive geometric estimation-type survival logic; they only change the proposal direction of a small number of candidates.

[0143] In embodiments of this disclosure, when the objective function is highly multimodal, the recent window cannot stably describe the current basin, or the curvature spectrum noise of the quadratic scalar model is large, the system sets w=0, retaining only development candidates. This mode reduces the dependence on inverse curvature exploration metrics and avoids local models misleading the main optimizer. w is gradually increased after the residual direction in the recent window is detected to be stable.

[0144] In the embodiments of this disclosure, in the industrial ISP image quality parameter tuning task, candidate point x represents a set of high-dimensional ISP configuration parameters. This configuration not only applies to the automatic white balance module, but can also simultaneously apply to multiple ISP or post-processing modules such as automatic exposure, brightness / contrast mapping, noise reduction, sharpening, color correction, tone mapping, local enhancement, dynamic range compression, and artifact suppression. The evaluation set can cover various scenes such as indoor, outdoor, night scene, backlight, low light, portrait, sky, greenery, text, and complex textures.

[0145] In the embodiments of this disclosure, for each candidate configuration, the system can calculate error or scoring components for multiple image quality dimensions, such as: noise residual, brightness deviation, local overexposure / underexposure, color cast, white balance error, skin tone deviation, saturation error, sharpening overshoot, texture fidelity reduction, insufficient dynamic range, artifact intensity, and differences in manual or model scoring under different scenes. These components can form a residual vector r(x), which is then converted into a residual s(x). Therefore, this embodiment does not protect a single automatic white balance parameter search, but rather a joint black-box parameter tuning framework encompassing "multiple scenes, multiple image quality dimensions, and multiple ISP modules."

[0146] In the embodiments of this disclosure, if the automatic white balance subtask includes 111 scene reproduction angle error optimization evaluation metrics, and the candidate configuration is 1135-dimensional ISP parameters, then each scene optimization evaluation metric can be used as part of the residual vector. Simultaneously, metrics such as noise, brightness, color, sharpness, and dynamic range can also be incorporated into the same vector as additional residual components. The system estimates the residual Jacobian J within the most recent approximately 300 evaluation windows, uses the side to generate damping candidates based on J and injects them into the main optimizer, and explores the side to construct the CMIX inverse curvature covariance based on the same J to perform trace-preserving shaping of the population sampling direction.

[0147] Corresponding to the data processing method described above, the present invention also proposes a data processing apparatus. Since the apparatus embodiments of the present invention correspond to the method embodiments described above, details not disclosed in the apparatus embodiments can be referred to in the method embodiments described above, and will not be repeated here.

[0148] Figure 7 This is a schematic diagram of the structure of a data processing apparatus provided in an embodiment of the present disclosure. The data processing apparatus includes: Acquisition unit 41 is used to acquire the first parameter generated by the parameter adjustment module in the target state; the first parameter is used by the data processing module to perform data processing; The determining unit 42 is used to determine the first evaluation result of the first parameter; the first evaluation result is used to evaluate the effect of data processing under the first parameter. The generation unit 43 is used to generate the second parameter based on the first evaluation result and the first parameter; The update unit 44 is used to update the target state based on the first evaluation result and the second parameter.

[0149] According to the data processing method proposed in this disclosure, the method includes obtaining a first parameter generated by a parameter adjustment module under a target state; the first parameter being used by a data processing module to perform data processing; determining a first evaluation result of the first parameter; the first evaluation result being used to evaluate the effect of data processing under the first parameter; generating a second parameter based on the first evaluation result and the first parameter; and updating the target state based on the first evaluation result and the second parameter. The second parameter is an additionally generated parameter. Updating the target state based on the first evaluation result and the second parameter enriches the features used to update the target state, thereby improving the update effect of the target state. This helps the parameter adjustment module generate better parameters, enabling the data processing module to use better parameters for data processing, indirectly improving the data processing effect of the data processing module.

[0150] In one possible implementation of this disclosure, the determining unit 42 is further configured to: Replace the original parameters in the data processing module with the first parameter. The original parameters are the parameters generated by the parameter adjustment module in the state before the target state. Based on the first parameter, the data processing module performs data processing on the first data to obtain the second data; Based on the second set of data, the first evaluation result is determined.

[0151] In one possible implementation of this disclosure, the determining unit 42 is further configured to: Get the labels of the first data; The first evaluation result is determined based on the difference between the second data and the label.

[0152] In one possible implementation of this disclosure, the generation unit 43 is further configured to: Evaluation indicators for obtaining the first evaluation result; According to the evaluation indicators, a set of differences is determined. The set of differences includes at least one difference. Each difference corresponds to one of the evaluation indicators. The difference is the difference between the label of the first data and the second data. The second data is obtained by the data processing module performing data processing on the first data under the first parameters. The second parameter is generated based on the difference set and the first parameter.

[0153] In one possible implementation of this disclosure, the generation unit 43 is further configured to: In response to a difference set including one difference, a virtual evaluation index is constructed based on the target state; Determine the virtual differences between the second data and the labels under the virtual evaluation index; The difference set is determined based on the evaluation indicators and the virtual evaluation indicators; the difference set includes the actual difference and the virtual difference.

[0154] In one possible implementation of this disclosure, the generation unit 43 is further configured to: The first parameter is normalized to obtain the normalized parameter; A correlation analysis was performed on the normalized parameters and the difference set to obtain the correlation matrix between the normalized parameters and the difference set; The second parameter is generated based on the correlation matrix.

[0155] In one possible implementation of this disclosure, the correlation matrix includes multiple correlation sub-matrices, the second parameter includes a third parameter and a fourth parameter, and the generation unit 43 is further configured to: Identify the target correlation submatrix with the highest correlation among multiple correlation submatrices; Parametric analysis is performed on the correlations in the target correlation submatrix to obtain the third parameter; A fourth parameter is obtained by performing parameter analysis on the correlation within a preset range; the preset range is the range based on the correlation matrix.

[0156] In one possible implementation of this disclosure, the updating unit 44 is further configured to: The second evaluation result is determined by the second parameter; The target status is updated based on the results of the first and second evaluations.

[0157] Since the apparatus provided in this embodiment corresponds to the methods provided in the above embodiments, the implementation of the methods is also applicable to the apparatus provided in this embodiment, and will not be described in detail in this embodiment.

[0158] The methods and apparatus provided in the embodiments of this application have been described above. To implement the functions of the methods provided in the embodiments of this application, the electronic device may include a hardware structure and software modules, and may implement the above functions in the form of a hardware structure, software modules, or a hardware structure plus software modules. One of the above functions may be executed in the form of a hardware structure, software modules, or a hardware structure plus software modules.

[0159] Figure 8 This is a block diagram illustrating an electronic device 500 for implementing the above-described data processing method according to an exemplary embodiment. For example, the electronic device 500 may be applied to servers, cloud environments, operational service platforms, various computer platforms, terminal systems, and web page systems.

[0160] Reference Figure 8The electronic device 500 may include one or more of the following components: processing component 502, memory 504, power supply component 506, multimedia component 508, audio component 510, input / output (I / O) interface 512, sensor component 514, and communication component 516.

[0161] Processing component 502 typically controls the overall operation of electronic device 500, such as operations associated with display, telephone calls, data communication, camera operation, and recording. Processing component 502 may include one or more processors 520 to execute instructions to perform all or part of the steps of the methods described above. Furthermore, processing component 502 may include one or more modules to facilitate interaction between processing component 502 and other components. For example, processing component 502 may include a multimedia module to facilitate interaction between multimedia component 508 and processing component 502.

[0162] Memory 504 is configured to store various types of data to support the operation of electronic device 500. Examples of this data include instructions for any application or method operating on electronic device 500, contact data, phonebook data, messages, pictures, videos, etc. Memory 504 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0163] Power supply component 506 provides power to various components of electronic device 500. Power supply component 506 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to electronic device 500.

[0164] Multimedia component 508 includes a screen that provides an output interface between electronic device 500 and a user. In this disclosure, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also the duration and pressure associated with the touch or swipe operation. In this disclosure, multimedia component 508 includes a front-facing camera and / or a rear-facing camera. When electronic device 500 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or rear-facing camera can receive external multimedia data. Each front-facing camera and rear-facing camera can be a fixed optical lens system or have focal length and optical zoom capabilities.

[0165] Audio component 510 is configured to output and / or input audio signals. For example, audio component 510 includes a microphone (MIC) configured to receive external audio signals when electronic device 500 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 504 or transmitted via communication component 516. In this disclosure, audio component 510 also includes a speaker for outputting audio signals.

[0166] I / O interface 512 provides an interface between processing component 502 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0167] Sensor assembly 514 includes one or more sensors for providing state assessments of various aspects of electronic device 500. For example, sensor assembly 514 can detect the on / off state of electronic device 500, the relative positioning of components such as the display and keypad of electronic device 500, changes in position of electronic device 500 or a component of electronic device 500, the presence or absence of user contact with electronic device 500, orientation or acceleration / deceleration of electronic device 500, and temperature changes of electronic device 500. Sensor assembly 514 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 514 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In this disclosure, sensor assembly 514 may also include an accelerometer, gyroscope, magnetometer, pressure sensor, or temperature sensor.

[0168] Communication component 516 is configured to facilitate wired or wireless communication between electronic device 500 and other devices. Electronic device 500 can access wireless networks based on communication standards, such as WiFi, 2G or 3G, 4G LTE, 5G NR (NewRadio), or combinations thereof. In one exemplary embodiment, communication component 516 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 516 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0169] In an exemplary embodiment, the electronic device 500 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0170] Embodiments of this disclosure also provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods described in the above embodiments of this disclosure.

[0171] Embodiments of this disclosure also provide a computer program product comprising a computer program executable by a programmable device, the computer program having, when executed by the programmable device, the method described in the above embodiments of this disclosure.

[0172] For cases where electronic devices can be chips or chip systems, see [link to relevant documentation]. Figure 9 The diagram shows the structure of the chip. Figure 9 The chip shown includes a processor 601 and an interface 602. There can be one or more processors 601, and multiple interfaces 602.

[0173] Optionally, the chip also includes a memory 603 for storing necessary computer programs and data.

[0174] Those skilled in the art will also understand that the various illustrative logical blocks and steps listed in the embodiments of this application can be implemented by electronic hardware, computer software, or a combination of both. Whether such functionality is implemented through hardware or software depends on the specific application and the overall system design requirements. Those skilled in the art can implement the functionality using various methods for each specific application, but such implementation should not be construed as exceeding the scope of protection of the embodiments of this application.

[0175] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0176] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0177] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0178] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processing module, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (control method), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic device, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0179] It should be understood that various parts of the embodiments of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0180] Those skilled in the art will understand that all or part of the steps of the methods described in the above embodiments can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium. When executed, the program includes one or a combination of the steps of the method embodiments.

[0181] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc.

[0182] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A data processing method, characterized in that, The method includes: The first parameter generated by the parameter adjustment module in the target state is obtained; the first parameter is used by the data processing module to perform data processing. Determine a first evaluation result for the first parameter; the first evaluation result is used to evaluate the effect of the data processing under the first parameter. Based on the first evaluation result and the first parameter, the second parameter is generated; The target state is updated based on the first evaluation result and the second parameter.

2. The method according to claim 1, characterized in that, The first evaluation result for determining the first parameter includes: Replace the original parameters in the data processing module with the first parameters, where the original parameters are those generated by the parameter adjustment module in the state before the target state. Based on the first parameter, the data processing module performs the data processing on the first data to obtain the second data; The first evaluation result is determined based on the second data.

3. The method according to claim 2, characterized in that, Determining the first evaluation result based on the second data includes: Obtain the tags of the first data; The first evaluation result is determined based on the difference between the second data and the label.

4. The method according to claim 1, characterized in that, The step of generating the second parameter based on the first evaluation result and the first parameter includes: Obtain the evaluation indicators for the first evaluation result; According to the evaluation index, a set of differences is determined. The set of differences includes at least one difference. The difference corresponds one-to-one with the evaluation index. The difference is the difference between the label of the first data and the second data. The second data is obtained by the data processing module performing the data processing on the first data under the first parameter. The second parameter is generated based on the set of differences and the first parameter.

5. The method according to claim 4, characterized in that, Determining the set of differences according to the evaluation indicators includes: In response to the inclusion of a difference in the set of differences, a virtual evaluation index is constructed based on the target state; Determine the virtual difference between the second data and the label under the virtual evaluation index; The difference set is determined according to the evaluation index and the virtual evaluation index; the difference set includes the difference and the virtual difference.

6. The method according to claim 4, characterized in that, The step of generating the second parameter based on the difference set and the first parameter includes: The first parameter is normalized to obtain the normalized parameter; A correlation analysis is performed on the normalized parameter and the difference set to obtain the correlation matrix between the normalized parameter and the difference set; The second parameter is generated based on the correlation matrix.

7. The method according to claim 6, characterized in that, The correlation matrix includes multiple correlation sub-matrices, and the second parameter includes a third parameter and a fourth parameter. Generating the second parameter based on the correlation matrix includes: Determine the target correlation submatrix with the highest correlation among the plurality of correlation submatrices; The third parameter is obtained by performing parametric analysis on the correlations in the target correlation submatrix. The fourth parameter is obtained by performing parameter analysis on the correlation within a preset range; the preset range is the range based on the correlation matrix.

8. The method according to claim 1, characterized in that, The step of updating the target state based on the first evaluation result and the second parameter includes: Determine the second evaluation result for the second parameter; The target state is updated based on the first evaluation result and the second evaluation result.

9. A data processing apparatus, characterized in that, The device includes: The acquisition unit is used to acquire the first parameter generated by the parameter adjustment module in the target state; the first parameter is used by the data processing module to perform data processing. A determining unit is configured to determine a first evaluation result for the first parameter; the first evaluation result is used to evaluate the effect of the data processing under the first parameter. The generation unit is used to generate the second parameter based on the first evaluation result and the first parameter; An update unit is used to update the target state based on the first evaluation result and the second parameter.

10. The apparatus according to claim 9, characterized in that, The determining unit is further configured to: Replace the original parameters in the data processing module with the first parameters, where the original parameters are those generated by the parameter adjustment module in the state before the target state. Based on the first parameter, the data processing module performs the data processing on the first data to obtain the second data; The first evaluation result is determined based on the second data.

11. The apparatus according to claim 9, characterized in that, The generation unit is also used for: Obtain the evaluation indicators for the first evaluation result; According to the evaluation index, a set of differences is determined. The set of differences includes at least one difference. The difference corresponds one-to-one with the evaluation index. The difference is the difference between the label of the first data and the second data. The second data is obtained by the data processing module performing the data processing on the first data under the first parameter. The second parameter is generated based on the set of differences and the first parameter.

12. The apparatus according to claim 9, characterized in that, The update unit is also used for: Determine the second evaluation result for the second parameter; The target state is updated based on the first evaluation result and the second evaluation result.

13. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-8.

14. A computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-8.

15. A program product, characterized in that, Includes computer instructions for causing a computer to perform the method of any one of claims 1-8.

Images