A monitoring and operation management method based on an artificial intelligence platform

Abstract

This invention provides a monitoring and operation management method based on an artificial intelligence platform, belonging to the field of operation management technology. Specifically, it includes: determining the model optimization target in a general model based on changes and accessed data; determining the similarity of training data between different matching business scenarios within the model optimization target; using the similarity to determine the similarity coefficient of training data between the matching business scenario and other matching business scenarios, and determining the optimization business scenario; obtaining the changes in training data within different time periods in different optimization business scenarios; and combining the similarity coefficient of training data in the optimization business scenario to determine the optimization target within the optimization business scenario; building a single artificial intelligence model for the optimization target, thereby improving the recognition accuracy of the artificial intelligence model.

Description

Technical Field

[0001] This invention belongs to the field of operations management technology, and in particular relates to a method for monitoring operations management based on an artificial intelligence platform. Background Technology

[0002] Artificial intelligence technology has good adaptability and flexibility in solving problems with complex nonlinearity, uncertainty, coupling and multivariable characteristics of power grid systems. It can play an important role in improving the operating efficiency, safety and reliability of power grids and their digitalization and intelligence levels.

[0003] By constructing an artificial intelligence platform, a development environment and ecosystem can be provided for different types of business needs, enabling better integration and collaboration with other technologies. Specifically, invention patent applications CN202210282690.9, "An Information Processing Method, System, and Cloud Platform Based on Artificial Intelligence," and CN202210322097.2, "Artificial Intelligence Service System and Method Based on Cloud Platform," both provide methods for building artificial intelligence cloud platforms in power grid enterprises. However, the existing technical solutions have the following technical problems:

[0004] Due to the complexity of the power grid's business needs, the artificial intelligence platform is configured with different types of artificial intelligence models based on different business needs. If there is a lack of statistics and analysis of the different artificial intelligence models on the artificial intelligence platform, users will not be able to effectively and quickly obtain and understand the overall situation of artificial intelligence. This makes it difficult to identify artificial intelligence models with low recognition accuracy or long recognition processing time, and thus it is difficult to guarantee the ease of use and accuracy of the artificial intelligence models on the platform.

[0005] To address the aforementioned technical problems, this invention provides a monitoring and operation management method based on an artificial intelligence platform. Summary of the Invention

[0006] To achieve the objectives of this invention, the following technical solution is adopted:

[0007] To address the aforementioned technical problems, the present invention provides the following technical solution to achieve its objectives:

[0008] A monitoring and operation management method based on an artificial intelligence platform, specifically including:

[0009] S1 uses the corresponding artificial intelligence models for different business scenarios as a basis to determine the general model and the matching business scenarios of the general model. When the recognition accuracy of the matching business scenario in the general model meets the requirements, proceed to the next step.

[0010] S2 acquires the call data and the changes in recognition accuracy under different matching business scenarios, and determines the model optimization target in the general model based on the changes and call data;

[0011] S3 determines the similarity of training data between different matching business scenarios in the model optimization objective, and uses the similarity to determine the similarity coefficient of training data between the matching business scenario and other matching business scenarios, as well as the optimization business scenario.

[0012] S4 acquires the changes in training data within different time periods in different optimization business scenarios, and determines the optimization target in the optimization business scenario by combining the similarity coefficient of the training data in the optimization business scenario, and builds a single artificial intelligence model for the optimization target.

[0013] The beneficial effects of this invention are as follows:

[0014] By determining whether the recognition accuracy of matching business scenarios in the general model meets the requirements, the system can filter matching business scenarios with low recognition accuracy using the general model. This allows for the creation of separate AI models for matching business scenarios, improving the recognition accuracy of AI models in matching business scenarios without increasing the management difficulty of the AI ​​models.

[0015] By analyzing the changes in training data across different time periods within various optimization business scenarios, and using the similarity coefficient of training data in each scenario to determine the optimization objectives, this approach effectively filters optimization objectives that might negatively impact the recognition accuracy of other matching business scenarios from both the similarity and volume of training data. This avoids the technical problem of using a general model where the recognition accuracy fails to meet requirements across multiple matching business scenarios. It not only ensures the recognition accuracy of the optimization objectives but also reduces the impact on the recognition accuracy of other matching business scenarios.

[0016] A further technical solution is that the general model is an artificial intelligence model adapted to multiple business scenarios.

[0017] A further technical solution is that the matching business scenario is a business scenario adapted to the general model.

[0018] A further technical solution involves determining that the recognition accuracy of the matching business scenario in the general model meets the requirements, specifically including:

[0019] Based on the recognition accuracy of the matching business scenario in the general model, the recognition accuracy under different business types in the matching business scenario is determined;

[0020] The identification deviation business type is determined based on the identification accuracy under different business types, and the identification accuracy of the matching business scenario in the general model is determined by the number of identification deviation business types.

[0021] A further technical solution is that when the recognition accuracy of the matching business scenario in the general model does not meet the requirements, a single artificial intelligence model is built for the matching business scenario.

[0022] A further technical solution is that the method for determining the optimization objective in the optimized business scenario is as follows:

[0023] The changes in training data across different time periods within various optimization scenarios are obtained. These changes are then combined with the similarity coefficients of the training data for each optimization scenario to determine the optimization objective. A single artificial intelligence model is then built for this optimization objective.

[0024] Based on the changes in training data in different time periods for the optimized business scenario, the proportion of training data in different time periods is determined, and the average proportion of training data in different time periods is used to determine the training data ratio coefficient for the optimized business scenario.

[0025] The model interference coefficient of the optimized business scenario is determined by multiplying the training data ratio coefficient and the training data similarity coefficient, and the optimization objective in the optimized business scenario is determined by using the model interference coefficient.

[0026] A further technical solution is that the value of the model interference coefficient is between 0 and 1, wherein the larger the model interference coefficient is, the greater the impact of the training data of the optimized business scenario on the recognition accuracy of other matching business scenarios.

[0027] Other features and advantages will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention are realized and obtained through the structures particularly pointed out in the description and the drawings.

[0028] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0029] The above and other features and advantages of the present invention will become more apparent from a detailed description of exemplary embodiments thereof with reference to the accompanying drawings.

[0030] Figure 1A flowchart illustrating a monitoring and operation management method based on an artificial intelligence platform;

[0031] Figure 2 This is a flowchart of a method for determining whether the recognition accuracy of a matching business scenario in the general model meets the requirements;

[0032] Figure 3 This is a flowchart illustrating the method for determining the model optimization objective in a general model;

[0033] Figure 4 It is a framework diagram of a computer system. Detailed Implementation

[0034] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, they are provided so that the invention will be thorough and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art. The same reference numerals in the drawings denote the same or similar structures, and therefore their detailed description will be omitted.

[0035] The terms “a,” “one,” “the,” and “the” are used to indicate the existence of one or more elements / components / etc.; the terms “including” and “having” are used to indicate an open-ended meaning of inclusion and that other elements / components / etc. may exist in addition to the listed elements / components / etc.

[0036] Example 1

[0037] To solve the above problems, according to one aspect of the present invention, such as Figure 1 As shown, according to one aspect of the present invention, a method for monitoring and managing operations based on an artificial intelligence platform is provided, specifically including:

[0038] S1 uses the corresponding artificial intelligence models for different business scenarios as a basis to determine the general model and the matching business scenarios of the general model. When the recognition accuracy of the matching business scenario in the general model meets the requirements, proceed to the next step.

[0039] S2 acquires the call data and the changes in recognition accuracy under different matching business scenarios, and determines the model optimization target in the general model based on the changes and call data;

[0040] S3 determines the similarity of training data between different matching business scenarios in the model optimization objective, and uses the similarity to determine the similarity coefficient of training data between the matching business scenario and other matching business scenarios, as well as the optimization business scenario.

[0041] S4 acquires the changes in training data within different time periods in different optimization business scenarios, and determines the optimization target in the optimization business scenario by combining the similarity coefficient of the training data in the optimization business scenario, and builds a single artificial intelligence model for the optimization target.

[0042] Furthermore, the general model is an artificial intelligence model adapted to multiple business scenarios.

[0043] Specifically, the matching business scenario refers to the business scenario adapted to the general model.

[0044] It should be noted that, as Figure 2 As shown, determining whether the recognition accuracy of the matching business scenario in the general model meets the requirements specifically includes:

[0045] Based on the recognition accuracy of the matching business scenario in the general model, the recognition accuracy under different business types in the matching business scenario is determined;

[0046] The identification deviation business type is determined based on the identification accuracy under different business types, and the identification accuracy of the matching business scenario in the general model is determined by the number of identification deviation business types.

[0047] Furthermore, when the recognition accuracy of the matching business scenario in the general model does not meet the requirements, a single artificial intelligence model is built for the matching business scenario.

[0048] Additionally, it should be noted that determining whether the recognition accuracy of the matching business scenario in the general model meets the requirements specifically includes:

[0049] Based on the recognition accuracy of the matching business scenario in the general model, the recognition accuracy under different business types in the matching business scenario is determined;

[0050] The accuracy of the matching business scenario in the general model is determined based on the average recognition accuracy under different business types.

[0051] Optionally, determining whether the recognition accuracy of the matching business scenario in the general model meets the requirements includes steps S11-S13, specifically:

[0052] S11 Based on the recognition accuracy of the matching business scenario in the general model, determine the recognition accuracy under different business types in the matching business scenario;

[0053] It should be noted that step S11 includes steps S111-S113, specifically:

[0054] S111 Based on the recognition accuracy of the matching business scenario in the general model, determine the recognition accuracy under different business types in the matching business scenario. When there is no business type whose recognition accuracy does not meet the requirements, directly determine that the recognition accuracy of the matching business scenario in the general model meets the requirements. When there is a business type whose recognition accuracy does not meet the requirements, proceed to step S112.

[0055] S112 identifies the business types whose recognition accuracy does not meet the requirements as the identification deviation business types. When the number of the identification deviation business types does not meet the requirements, it is determined that the recognition accuracy of the matching business scenario in the general model does not meet the requirements. When the number of the identification deviation business types meets the requirements, proceed to step S113.

[0056] S113 determines the recognition deviation coefficient of the matching business scenario based on the recognition accuracy of different recognition deviation types. When the recognition deviation coefficient of the matching business scenario does not meet the requirements, it is determined that the recognition accuracy of the matching business scenario in the general model does not meet the requirements. When the recognition deviation coefficient of the matching business scenario meets the requirements, proceed to step S12.

[0057] S12 determines the number of calls under different business types based on the call data under different business types, and determines the application reliability coefficient under different business types based on the number of calls and the recognition accuracy.

[0058] It should be further explained that step S12 above includes S121-S123, specifically:

[0059] S121 determines the number of calls under different business types based on the call data under different business types, and takes the business types whose recognition accuracy does not meet the requirements as the recognition deviation business types. When the total number of calls under the recognition deviation business types does not meet the requirements, it is determined that the recognition accuracy of the matching business scenario in the general model does not meet the requirements. When the total number of calls under the recognition deviation business types meets the requirements, proceed to step S122.

[0060] S122 determines the application reliability coefficient under different business types based on the number of calls and the recognition accuracy. When there is a business type whose application reliability coefficient does not meet the requirements, proceed to step S123. When there is no business type whose application reliability coefficient does not meet the requirements, it is determined that the recognition accuracy of the matching business scenario in the general model meets the requirements.

[0061] S123 When the number of service types whose application reliability coefficient does not meet the requirements does not meet the requirements, it is determined that the recognition accuracy of the matching service scenario in the general model does not meet the requirements. When the number of service types whose application reliability coefficient does not meet the requirements meets the requirements, proceed to step S13.

[0062] S13 determines the comprehensive reliability coefficient based on the application reliability coefficient under different business types, and determines whether the recognition accuracy of the matching business scenario in the general model meets the requirements based on the comprehensive reliability coefficient.

[0063] Furthermore, the call data includes the historical call count of the general model under the matching business scenario.

[0064] It is understood that the variation in recognition accuracy is determined based on the variation in recognition accuracy between different time periods.

[0065] Specifically, the time periods are divided according to the training and update times of the general model.

[0066] Specifically, such as Figure 3 As shown, the method for determining the model optimization objective in the general model is as follows:

[0067] The total number of historical calls to the general model is determined based on the call data under different matching business scenarios, and the call frequency coefficient of the general model is determined through the total number of historical calls.

[0068] Based on the variation of recognition accuracy under different matching business scenarios, the matching business scenarios where recognition accuracy deteriorates are identified and these are taken as recognition degradation scenarios. The call deviation coefficient of the general model is determined by the number of recognition degradation scenarios.

[0069] The model optimization requirement coefficient of the general model is determined by using the call deviation coefficient and the call frequency coefficient, and whether the general model is the model optimization target is determined based on the model optimization requirement coefficient.

[0070] Furthermore, the model optimization requirement coefficient is determined based on the average of the call deviation coefficient and the call frequency coefficient.

[0071] It is understood that when the model optimization demand coefficient of the general model is greater than the preset demand coefficient threshold, the general model is determined as the model optimization target.

[0072] Specifically, the similarity of the training data includes the similarity of the input features and the similarity of the output values.

[0073] Optionally, the method for determining the optimized business scenario is as follows:

[0074] The similarity between the training data of the matching business scenario and other matching business scenarios is used to determine the similarity between the training data of the matching business scenario and other matching business scenarios under different business types, and the similarity between the input features and output quantities of the training data is determined based on the similarity.

[0075] The similarity between the input features and the output quantities is used to determine the similarity of the training data with other matching business scenarios under different business types, and the similar business types are determined based on the similarity.

[0076] The similarity coefficient between the matching business scenario and other matching business scenarios is determined based on the number of similar business types, and the optimized business scenario in the matching business scenario is determined based on the similarity coefficient of the training data.

[0077] It should be noted that when the similarity meets the requirements, the business type is determined to be a similar business type.

[0078] It should be noted that when the similarity coefficient of the training data is greater than the preset similarity coefficient threshold, the matching business scenario is determined to be an optimized business scenario.

[0079] Optionally, the determination of the optimized business scenario includes steps S31-S33, specifically:

[0080] S31 determines the similarity between the training data of the matching business scenario and other matching business scenarios under different business types based on the similarity between the training data of the matching business scenario and other matching business scenarios, and determines the similarity between the input features and the output of the training data based on the similarity.

[0081] Optionally, before proceeding to step S32, it is also necessary to determine whether there are any business types in which the similarity of input features between the training data of the matching business scenario and other matching business scenarios under different business types does not meet the requirements. If there are no business types in which the similarity of input features does not meet the requirements, it is determined that the matching business scenario does not belong to the optimized business scenario. If and only if there are business types in which the similarity of input features does not meet the requirements, proceed to step S32.

[0082] S32 determines the similarity of training data with other matching business scenarios under different business types by the similarity of the input features and the similarity of the output quantities, and determines the similar business types based on the similarity.

[0083] Optionally, step S32 above includes steps S321-S322, specifically as follows:

[0084] S321 determines the similarity of the training data with other matching business scenarios under different business types by the similarity of the input features and the similarity of the output. When there is no business type whose similarity does not meet the requirements, it is determined that the matching business scenario does not belong to the optimized business scenario. When there is a business type whose similarity does not meet the requirements, proceed to step S322.

[0085] S322 If there is a business type with a similarity greater than the preset similarity, then the matched business scenario is determined to be an optimized business scenario; if there is no business type with a similarity greater than the preset similarity, then proceed to step S323.

[0086] S323 determines similar business types based on the similarity. If the number of similar business types does not meet the requirements, the matching business scenario is determined to be an optimized business scenario. If the number of similar business types meets the requirements, the process proceeds to step S33.

[0087] S33 determines the training data similarity coefficient between the matched business scenario and other matched business scenarios based on the number of similar business types and the similarity between different similar business types, and determines the optimized business scenario in the matched business scenario based on the training data similarity coefficient.

[0088] Specifically, the method for determining the optimization objective in the aforementioned business scenario is as follows:

[0089] The changes in training data across different time periods within various optimization scenarios are obtained. These changes are then combined with the similarity coefficients of the training data for each optimization scenario to determine the optimization objective. A single artificial intelligence model is then built for this optimization objective.

[0090] Based on the changes in training data in different time periods for the optimized business scenario, the proportion of training data in different time periods is determined, and the average proportion of training data in different time periods is used to determine the training data ratio coefficient for the optimized business scenario.

[0091] The model interference coefficient of the optimized business scenario is determined by multiplying the training data ratio coefficient and the training data similarity coefficient, and the optimization objective in the optimized business scenario is determined by using the model interference coefficient.

[0092] A further technical solution is that the value of the model interference coefficient is between 0 and 1, wherein the larger the model interference coefficient is, the greater the impact of the training data of the optimized business scenario on the recognition accuracy of other matching business scenarios.

[0093] Example 2

[0094] Secondly, such as Figure 4 As shown, the present invention provides a computer system, including: a memory and a processor connected in communication, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the above-described monitoring and operation management method based on an artificial intelligence platform when running the computer program.

[0095] Optionally, the method for determining the model optimization objective in the general model is as follows:

[0096] S21 determines the number of times the general model is called in different matching business scenarios based on the call data in different matching business scenarios, and determines the call frequency coefficient of the general model by the number of times it is called in different matching business scenarios;

[0097] S22 determines the matching business scenarios where the recognition accuracy deteriorates based on the changes in recognition accuracy under different matching business scenarios, and takes these scenarios as recognition degradation scenarios. The general model's call deviation coefficient is determined by the number of recognition degradation scenarios, the number of calls to recognition degradation scenarios, and the change in recognition accuracy.

[0098] S23 uses the call deviation coefficient and the call frequency coefficient to determine the model optimization requirement coefficient of the general model, and determines whether the general model is the model optimization target based on the model optimization requirement coefficient.

[0099] Optionally, step S21 above includes steps S211-S213, specifically as follows:

[0100] S211 determines the number of times the general model is called in different matching business scenarios based on the call data in different matching business scenarios. When the number of matching business scenarios with a call count greater than the preset number of calls is greater than the preset number of scenarios, the general model is determined as the model optimization target. When the number of matching business scenarios with a call count greater than the preset number of calls is not greater than the preset number of scenarios, proceed to step S222.

[0101] S222 determines the total number of calls to the general model by counting the number of calls under different matching business scenarios. When the total number of calls to the general model is greater than a preset call count threshold, the general model is determined as the model optimization target. When the total number of calls to the general model is not greater than the preset call count threshold, proceed to step S223.

[0102] S223 determines the call frequency coefficient of the general model by the number of calls under different matching business scenarios. When the call frequency coefficient of the general model is greater than the preset frequency coefficient threshold, the general model is determined as the model optimization target. When the call frequency coefficient of the general model is not greater than the preset frequency coefficient threshold, proceed to step S23.

[0103] Optionally, step S22 above includes steps S221-S222, specifically as follows:

[0104] S221 Based on the changes in recognition accuracy under different matching business scenarios, identify the matching business scenarios where the recognition accuracy deteriorates and take them as recognition degradation scenarios. When the number of recognition degradation scenarios does not meet the requirements, the general model is determined as the model optimization target. When the number of recognition degradation scenarios meets the requirements, proceed to step S222.

[0105] S222 determines the calling deviation coefficient of the general model by the number of identified deteriorated scenarios, the number of times the identified deteriorated scenarios are called, and the change in the recognition accuracy. When the calling deviation coefficient of the general model does not meet the requirements, the general model is determined as the model optimization target. When the calling deviation coefficient of the general model meets the requirements, proceed to step S23.

[0106] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0107] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0108] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A monitoring and operation management method based on an artificial intelligence platform, characterized in that, Specifically, it includes: Based on the corresponding artificial intelligence models in different business scenarios, a general model and the matching business scenarios of the general model are determined. When the recognition accuracy of the matching business scenario in the general model meets the requirements, proceed to the next step. Obtain call data and changes in recognition accuracy under different matching business scenarios, and determine the model optimization target in the general model based on the changes and call data; Determine the similarity of training data between different matching business scenarios in the model optimization objective, and use the similarity to determine the similarity coefficient of training data between the matching business scenario and other matching business scenarios, as well as the optimization business scenario. The changes in training data within different time periods in different optimization business scenarios are obtained, and the optimization objectives in the optimization business scenarios are determined by combining the similarity coefficient of the training data in the optimization business scenarios, and a single artificial intelligence model is built for the optimization objectives.

2. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, The general model is an artificial intelligence model adapted to multiple business scenarios.

3. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, The matching business scenario refers to the business scenario adapted to the general model.

4. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, Determining whether the recognition accuracy of the matching business scenario in the general model meets the requirements specifically includes: Based on the recognition accuracy of the matching business scenario in the general model, the recognition accuracy under different business types in the matching business scenario is determined; The identification deviation business type is determined based on the identification accuracy under different business types, and the identification accuracy of the matching business scenario in the general model is determined by the number of identification deviation business types.

5. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, When the recognition accuracy of the matching business scenario in the general model does not meet the requirements, a single artificial intelligence model is built for the matching business scenario.

6. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, The call data includes the historical call count of the general model under the matching business scenario.

7. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, The variation in recognition accuracy is determined based on the amount of variation in recognition accuracy between different time periods.

8. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, The time periods are divided based on the training and update times of the general model.

9. The monitoring and operation management method based on an artificial intelligence platform as described in claim 1, characterized in that, The method for determining the model optimization objective in the general model is as follows: The total number of historical calls to the general model is determined based on the call data under different matching business scenarios, and the call frequency coefficient of the general model is determined through the total number of historical calls. Based on the variation of recognition accuracy under different matching business scenarios, the matching business scenarios where recognition accuracy deteriorates are identified and these are taken as recognition degradation scenarios. The call deviation coefficient of the general model is determined by the number of recognition degradation scenarios. The model optimization requirement coefficient of the general model is determined by using the call deviation coefficient and the call frequency coefficient, and whether the general model is the model optimization target is determined based on the model optimization requirement coefficient.

10. The monitoring and operation management method based on an artificial intelligence platform as described in claim 9, characterized in that, The model optimization requirement coefficient is determined based on the average of the call deviation coefficient and the call frequency coefficient.

Images