Data processing method and device, storage medium and electronic equipment

Abstract

The invention discloses a data processing method and device, a storage medium and electronic equipment. Relates to the technical field of data processing, the field of automation control and other related technical fields. The method comprises the following steps: acquiring a data processing result and target parameter information of a target data processing request triggered by an artificial intelligence decision-making system for a target object; identifying the target parameter information to obtain an interpretation text corresponding to the data processing result; and feeding back the interpretation text and the data processing result to the target object. Through the data processing method and apparatus, the problem of poor interpretability of the data processing result caused by black box operation when the AI decision system processes the data processing request in the related technology is solved.

Description

Technical Field

[0001] The present application relates to the fields of data processing technology, automation control and other related technical fields, and more specifically, to a data processing method and device, a storage medium and an electronic device. Background Art

[0002] Traditional AI systems rely on complex algorithms and large amounts of data input to provide deep insights and decision support. However, AI systems operate in a black box when processing data processing requests, and the decision-making process is usually opaque and difficult for non-professionals to understand, limiting the widespread application of AI technology in industries that have strict requirements for transparency and explainability. In addition, as AI technology is increasingly used in automated tasks, the need for review and supervision of AI decision-making processes is also increasing. Robotic Process Automation (RPA) technology can automate repetitive tasks by simulating user operations, but RPA technology still has significant limitations in handling complex decisions and providing decision explanations.

[0003] Currently, no effective solution has been proposed to the problem that the AI ​​decision-making system in related technologies processes data processing requests as a black box operation, resulting in poor interpretability of data processing results. Summary of the invention

[0004] The main purpose of this application is to provide a data processing method and device, a storage medium and an electronic device to solve the problem in the related art that the AI ​​decision-making system performs black box operations when processing data processing requests, resulting in poor interpretability of data processing results.

[0005] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a data processing method is provided. The method comprises: obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision system when processing the target data processing request; identifying the target parameter information to obtain the explanation text corresponding to the data processing result; and feeding back the explanation text and the data processing result to the target object.

[0006] Furthermore, obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system for the target object includes: parsing and processing the target data processing request by the data preprocessing unit of the artificial intelligence decision system to obtain target request data, wherein the target request data at least includes: the request time, request content and request context information of the target data processing request; performing feature extraction on the target request data by the feature extraction unit of the artificial intelligence decision system to obtain target feature data; and processing the target data processing request based on the target feature data by the artificial intelligence decision system to obtain the data processing result and the target parameter information.

[0007] Furthermore, the target parameter information is identified to obtain an explanatory text corresponding to the data processing result, including: performing word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; performing vectorization processing on the target parameter information after word segmentation to obtain the processed target parameter information; and identifying the processed target parameter information to obtain the explanatory text.

[0008] Furthermore, feeding back the explanatory text and the data processing result to the target object includes: judging whether the explanatory text complies with a preset execution logic, and whether the data processing result complies with an expected result, and obtaining a judgment result; if the judgment result indicates that the explanatory text complies with the preset execution logic, and the data processing result complies with the expected result, then visualizing the explanatory text and the data processing result to feed back the explanatory text and the data processing result to the target object.

[0009] Furthermore, after determining whether the explanatory text complies with the preset execution logic and whether the data processing result complies with the expected result and obtaining the judgment result, the method also includes: if the judgment result indicates that the explanatory text does not comply with the preset execution logic, or the data processing result does not comply with the expected result, determining the deviation level corresponding to the explanatory text and the data processing result; determining the target adjustment strategy based on the deviation level; adjusting the artificial intelligence decision-making system based on the target adjustment strategy, and obtaining the data processing result and the explanatory text again based on the adjusted artificial intelligence decision-making system.

[0010] Furthermore, the target data processing request is parsed and processed by the data preprocessing unit of the artificial intelligence decision-making system to obtain the target request data, including: parsing and processing the target data processing request to obtain initial request data; cleaning the initial request data to obtain cleaned initial request data; and filling in the missing parts of the cleaned initial request data to obtain the target request data.

[0011] Furthermore, the feature extraction unit of the artificial intelligence decision-making system performs feature extraction on the target request data to obtain target feature data, including: performing feature extraction on the target request data through the convolution layer in the feature extraction unit to obtain initial feature data; performing dimensionality reduction processing on the initial feature data through the pooling layer in the feature extraction unit to obtain initial feature data after dimensionality reduction; and outputting the initial feature data after dimensionality reduction through the output layer in the feature extraction unit to obtain the target feature data.

[0012] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a data processing device is provided. The device comprises: an acquisition unit, used to acquire the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision system when processing the target data processing request; an identification unit, used to identify the target parameter information and obtain the explanation text corresponding to the data processing result; and a feedback unit, used to feed back the explanation text and the data processing result to the target object.

[0013] Furthermore, the acquisition unit includes: a parsing module, used to parse and process the target data processing request through the data preprocessing unit of the artificial intelligence decision-making system to obtain target request data, wherein the target request data at least includes: the request time, request content and request context information of the target data processing request; a feature extraction module, used to extract features from the target request data through the feature extraction unit of the artificial intelligence decision-making system to obtain target feature data; a first processing module, used to process the target data processing request based on the target feature data through the artificial intelligence decision-making system to obtain the data processing result and the target parameter information.

[0014] Furthermore, the recognition unit includes: a second processing module, used to perform word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; a third processing module, used to perform vectorization processing on the target parameter information after word segmentation to obtain the processed target parameter information; and a recognition module, used to recognize the processed target parameter information to obtain the explanation text.

[0015] Furthermore, the feedback unit includes: a judgment module, used to judge whether the interpretation text complies with the preset execution logic, and whether the data processing result complies with the expected result, to obtain a judgment result; a fourth processing module, used to visualize the interpretation text and the data processing result if the judgment result indicates that the interpretation text complies with the preset execution logic and the data processing result complies with the expected result, so as to feed back the interpretation text and the data processing result to the target object.

[0016] Furthermore, the device also includes: a first determination unit, which is used to determine whether the explanatory text conforms to the preset execution logic and whether the data processing result conforms to the expected result. After obtaining the judgment result, if the judgment result indicates that the explanatory text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, then determine the deviation level corresponding to the explanatory text and the data processing result; a second determination unit, which is used to determine a target adjustment strategy based on the deviation level; and an adjustment unit, which is used to adjust the artificial intelligence decision-making system based on the target adjustment strategy, and obtain the data processing result and the explanatory text again based on the adjusted artificial intelligence decision-making system.

[0017] Furthermore, the parsing module includes: a parsing submodule, which is used to parse the target data processing request to obtain initial request data; a cleaning submodule, which is used to clean the initial request data to obtain cleaned initial request data; and a filling submodule, which is used to fill in the missing parts of the cleaned initial request data to obtain the target request data.

[0018] Furthermore, the feature extraction module includes: a feature extraction submodule, used to perform feature extraction on the target request data through the convolution layer in the feature extraction unit to obtain initial feature data; a dimension reduction submodule, used to perform dimension reduction processing on the initial feature data through the pooling layer in the feature extraction unit to obtain the initial feature data after dimension reduction; and an output submodule, used to output the initial feature data after dimension reduction through the output layer in the feature extraction unit to obtain the target feature data.

[0019] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a computer-readable storage medium is provided, wherein the storage medium stores a program, wherein when the program is running, the device where the storage medium is located is controlled to execute any one of the data processing methods described above.

[0020] In order to achieve the above-mentioned purpose, according to another aspect of the present application, an electronic device is further provided, the electronic device includes one or more processors and a memory, and the memory is used to store the data processing method implemented by one or more processors as described above.

[0021] Through this application, the following steps are adopted: obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision system when processing the target data processing request; identifying the target parameter information to obtain the explanatory text corresponding to the data processing result; feeding back the explanatory text and the data processing result to the target object. Through this application, the problem that the AI ​​decision system in the related art is a black box operation when processing the data processing request, resulting in poor interpretability of the data processing result is solved. In this scheme, according to the data processing result after the artificial intelligence decision system processes the data processing request triggered by the target object and the target parameter information called when processing the target data processing request, an explanatory text corresponding to the data processing result is generated, and the explanatory text and the data processing result are synchronously fed back to the target object. Through the data processing result and the target parameter information, a complete track of the decision created by the artificial intelligence decision system is created, so that the decision process of the artificial intelligence decision system is traceable and reviewable. At the same time, the explanatory text corresponding to the data processing result is fed back to the target object to help the target object understand the decision reasons of the artificial intelligence decision system and enhance the transparency of the artificial intelligence decision system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0023] Figure 1 is a flow chart of a data processing method provided according to an embodiment of the present application;

[0024] Figure 2 is a schematic diagram of a data processing method provided according to an embodiment of the present application;

[0025] Figure 3 is a schematic diagram of a data processing device provided according to an embodiment of the present application;

[0026] Figure 4 is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0028] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties. For example, an interface is set between this system and the relevant user or organization. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0031] The present invention is described below in conjunction with preferred implementation steps. Figure 1 is a flow chart of a data processing method provided according to an embodiment of the present application, such as Figure 1 As shown, the method comprises the following steps:

[0032] Step S101, obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision-making system on the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision-making system when processing the target data processing request.

[0033] Optionally, the target data processing requests triggered by the target object in the ERP system, sales records and other business systems are collected through the data collector unit, and the target parameter information such as all input parameters, parameter settings and intermediate calculation results used by the AI ​​model (i.e., the artificial intelligence decision-making system) when making decisions (i.e., when processing the target data processing requests) are recorded through the tracking and recording unit.

[0034] It should be noted that the tracking and recording unit can ensure that every step of the decision-making process is recorded in detail. By capturing the decision logic and basis of the AI ​​model, the transparency of AI decision-making is enhanced, so that the target object can understand how the AI ​​model reaches a conclusion.

[0035] Step S102, identifying the target parameter information and obtaining an explanation text corresponding to the data processing result.

[0036] Optionally, the natural language processing (NLP) technology in the interpretation algorithm unit is used to convert the complex decision logic of the AI ​​model into a form that is easy for the target object to understand based on the target parameter information to obtain an interpretation text. The content of the interpretation text includes but is not limited to: the processing reasons and processing basis of the AI ​​model when processing target data processing requests, etc.

[0037] Step S103, feeding back the interpretation text and data processing results to the target object.

[0038] Optionally, the decision-making process and results of AI (i.e., explanation text and data processing results) are graphically displayed through a visualization unit and presented to the user (i.e., the target object) through a dashboard.

[0039] In summary, the data processing results after the artificial intelligence decision system processes the data processing request triggered by the target object and the target parameter information called when processing the target data processing request are identified, an explanation text corresponding to the data processing result is generated, and the explanation text and the data processing result are synchronously fed back to the target object. Through the data processing results and target parameter information, a complete track of the artificial intelligence decision system creating decisions is created, making the decision process of the artificial intelligence decision system traceable and reviewable. At the same time, the explanation text corresponding to the data processing result is fed back to the target object to help the target object understand the decision reasons of the artificial intelligence decision system, thereby enhancing the transparency of the artificial intelligence decision system.

[0040] Optionally, in the data processing method provided in the embodiment of the present application, obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object includes: parsing and processing the target data processing request by the data preprocessing unit of the artificial intelligence decision system to obtain target request data, wherein the target request data includes at least: the request time, request content and request context information of the target data processing request; performing feature extraction on the target request data by the feature extraction unit of the artificial intelligence decision system to obtain target feature data; processing the target data processing request based on the target feature data by the artificial intelligence decision system to obtain data processing results and target parameter information.

[0041] In an optional embodiment, the target data processing request triggered by the target object in the ERP system, sales records and other business systems is collected by the data collector unit, and parsed to obtain the target request data such as the request time, request content and request context information about the target data processing request. Then, the feature extraction unit extracts key information from the target request data for performance analysis and anomaly detection, and screens out the most important data features to obtain the target feature data. Finally, the target data processing request is processed based on the target feature data by the artificial intelligence decision system to obtain the data processing result, and the tracking and recording unit is used to record the target parameter information such as all input parameters, parameter settings and intermediate calculation results used by the AI ​​model (i.e., the artificial intelligence decision system) when making decisions (i.e., when processing the target data processing request).

[0042] Through multiple functional units such as the data collector unit and the feature extraction unit, the data processing results and target parameter information of the target data processing request triggered by the artificial intelligence decision-making system on the target object can be obtained quickly and accurately, thereby improving the target object's trust in the artificial intelligence decision-making system.

[0043] Optionally, in the data processing method provided in the embodiment of the present application, the target parameter information is identified to obtain an explanatory text corresponding to the data processing result, including: performing word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; performing vectorization processing on the target parameter information after word segmentation to obtain the processed target parameter information; and identifying the processed target parameter information to obtain an explanatory text.

[0044] In an optional embodiment, the target parameter information is decomposed into smaller units based on the word segmenter by the natural language processing technology (NLP) in the interpretation algorithm unit to obtain the target parameter information after word segmentation. The target parameter information after word segmentation is converted into a numerical form using a vectorization tool to obtain the processed target parameter information. Finally, the interpretation text corresponding to the data processing result is identified based on the processed target parameter information.

[0045] It should be noted that vectorization tools include but are not limited to: Bag of Words, TF-IDF (Term Frequency-Inverse Document Frequency), Word2Vec, BERT and other tools.

[0046] It should be noted that, for example, if a manufacturing company uses AI models to predict product demand and optimize inventory management, the AI ​​model predicts that the demand for a certain product will increase by 20% next quarter. The explanation algorithm unit will analyze the logic and data basis behind this prediction result, such as identifying possible factors that lead to increased demand (for example, holiday impact, market promotions, or industry trends, etc.). The unit then converts these technical details into easy-to-understand explanation text, such as: "According to analysis, it is expected that demand will increase by 20% next quarter due to the promotion of a certain festival, combined with the current market's positive feedback on this product category."

[0047] It should be noted that, for example, a manufacturing company uses an AI model to manage inventory: the AI ​​model may predict a shortage of inventory for a specific product. The explanation algorithm unit will provide a detailed explanation of the cause of this prediction, such as: "expected growth due to supply chain disruptions, sudden increase in demand, or new styles of products to be released soon."

[0048] It should be noted that, for example, a manufacturing company uses an AI model to predict product demand in different regions or markets: the AI ​​model will analyze historical sales data, market trends, seasonal factors, etc. The explanation algorithm unit will analyze the prediction results of the AI ​​model and provide explanations for key factors, such as: "The increase in demand for a certain product is due to the approaching of a specific holiday or the impact of a certain sporting event."

[0049] By interpreting the algorithm unit, the complex decision logic of the AI ​​model is converted into a form that is easy for users to understand, which helps the target object understand the AI ​​decision-making process, enhances the target object's trust in AI decisions, and helps relevant departments of the enterprise formulate corresponding processing strategies.

[0050] Optionally, in the data processing method provided in the embodiment of the present application, feeding back the interpretation text and the data processing results to the target object includes: judging whether the interpretation text conforms to the preset execution logic, and whether the data processing results conform to the expected results, and obtaining a judgment result; if the judgment result characterizes that the interpretation text conforms to the preset execution logic, and the data processing results conform to the expected results, then visualizing the interpretation text and the data processing results to feed back the interpretation text and the data processing results to the target object.

[0051] In an optional embodiment, the compliance checking unit automatically checks whether the interpretation text of the AI ​​model complies with the preset execution logic. If the interpretation text does not comply with the preset execution logic, it means that the output of the interpretation text is incorrect and needs to be regenerated according to the preset execution logic. The compliance checking unit automatically checks whether the output results of the AI ​​model comply with the preset compliance rules (i.e., whether the data processing results meet the expected results). Take the example of a manufacturing company predicting product demand through an AI model: the compliance checking unit will verify whether the prediction results of the output of the AI ​​model meet the expected results (for example: whether the output prediction results are within a reasonable inventory level range, whether the output prediction results exceed the inventory upper limit and budget limit, whether the output prediction results comply with the company's inventory policy and procurement regulations, etc.). If the output prediction results exceed these limits, the compliance checking unit will issue a warning and make adjustment suggestions (for example: reduce the forecast amount or find a more cost-effective procurement channel). If the explanation text conforms to the preset execution logic and the data processing result conforms to the expected result, the explanation text and the data processing result are displayed in a graphical manner (for example, a chart, a flowchart or a decision tree, etc.) through a visualization display unit to feed back the explanation text and the data processing result to the target object.

[0052] It should be noted that in order to help managers intuitively understand the forecast results output by the AI ​​model (i.e., explain the text and data processing results), the visualization unit will create charts and graphs, such as a line chart showing historical sales data and forecast demand curves or a bar chart showing the impact of different factors on demand forecasts. In addition, if the forecast results output by the AI ​​model decrease, the unit can display a chart with warning colors to intuitively display the potential risk of inventory backlogs.

[0053] Through visualization methods, users can intuitively see how the data is processed and how the AI ​​model makes the final decision, further enhancing the transparency of decision-making.

[0054] Optionally, in the data processing method provided in the embodiment of the present application, after determining whether the explanation text conforms to the preset execution logic and whether the data processing result conforms to the expected result, and obtaining the judgment result, the method further includes: if the judgment result indicates that the explanation text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, determining the deviation level corresponding to the explanation text and the data processing result; determining the target adjustment strategy based on the deviation level; adjusting the artificial intelligence decision-making system based on the target adjustment strategy, and obtaining the data processing result and the explanation text again based on the adjusted artificial intelligence decision-making system.

[0055] In an optional embodiment, if the judgment result indicates that the interpretation text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, the risk assessment unit analyzes the deviation level corresponding to the interpretation text and the data processing result, and determines the possible risks (for example, inventory backlog risk or out-of-stock risk, etc.) according to the deviation level, and evaluates the potential impact of these risks on the operation of the enterprise to determine the target adjustment strategy. Finally, the AI ​​model is adjusted according to the target adjustment strategy to obtain the data processing result and the interpretation text again.

[0056] It should be noted that if the data processing result of the AI ​​model is "the demand for a certain product will increase significantly", the risk assessment unit will evaluate the pressure of this surge in demand on the supply chain, as well as the risks such as insufficient or oversupply that may result.

[0057] It should be noted that the Alert Manager unit is used to monitor the operation of the AI ​​model and automatically generate alerts when potential problems or deviations are detected. For example, if the AI ​​model predicts that the inventory of a high-demand product is about to run out, the dashboard display unit will highlight this information and send it to inventory managers in real time by the Alert Manager unit, prompting them to take action, such as speeding up order processing or adjusting production priorities.

[0058] By judging whether the explanatory text conforms to the preset execution logic and whether the data processing results meet the expected results, it is possible to quickly and accurately determine whether the data processing results and explanatory text output by the AI ​​model need to be adjusted, thereby improving the decision-making accuracy of the AI ​​model.

[0059] Optionally, in the data processing method provided in the embodiment of the present application, the target data processing request is parsed and processed by the data preprocessing unit of the artificial intelligence decision-making system to obtain the target request data, including: parsing the target data processing request to obtain initial request data; cleaning the initial request data to obtain cleaned initial request data; and filling in the missing parts of the cleaned initial request data to obtain the target request data.

[0060] In an optional embodiment, the data preprocessing unit directly performs necessary preprocessing operations on the original data (i.e., the initial request data) collected by the data collector unit, and the preprocessing operations include but are not limited to: cleaning processing of the initial request data and filling in the gaps of the initial request data, so as to obtain the target request data.

[0061] It should be noted that the cleaning process of the initial request data is to clear the erroneous data (eliminate the random errors or invalid information in the data). The missing filling process of the initial request data is to process the blanks or missing parts in the data.

[0062] The data preprocessing unit of the artificial intelligence decision-making system can quickly parse, clean and fill in the gaps in the target data processing request, thereby accurately obtaining the target request data.

[0063] Optionally, in the data processing method provided in the embodiment of the present application, feature extraction is performed on the target request data through the feature extraction unit of the artificial intelligence decision-making system to obtain the target feature data, including: feature extraction is performed on the target request data through the convolution layer in the feature extraction unit to obtain initial feature data; dimension reduction is performed on the initial feature data through the pooling layer in the feature extraction unit to obtain initial feature data after dimension reduction; and the initial feature data after dimension reduction is output through the output layer in the feature extraction unit to obtain target feature data.

[0064] In an optional embodiment, the input data (i.e., the target request data) is feature extracted by a convolution layer in a feature extraction unit: a set of learnable filters (convolution kernels) are used to slide on the input data, and the weighted sum of the coverage area is calculated to obtain initial feature data. Then, the spatial dimension of the initial feature data is reduced by a pooling layer in the feature extraction unit (usually using maximum pooling or average pooling) to obtain the initial feature data after dimensionality reduction. Finally, the initial feature data after dimensionality reduction is converted and output by an output layer in the feature extraction unit to obtain the target feature data.

[0065] Through the convolution layer, pooling layer and output layer, the target request data can be quickly processed by feature extraction, dimensionality reduction and data conversion, thereby accurately achieving efficient data processing and learning.

[0066] It should be noted that Figure 2 Schematic diagram of a data processing method according to an embodiment of the present application. Figure 2 As shown in the figure, the data processing method consists of multiple functional modules, including but not limited to: data collection and preprocessing module, AI decision tracking module, decision explanation system, user interaction interface and compliance review mechanism. By combining the automated execution capability of RPA and the advanced analysis capability of AI, the transparency and explainability of AI decisions are improved, user trust is enhanced, and the compliance of AI decisions is ensured.

[0067] The data processing method provided in the embodiment of the present application obtains the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision system when processing the target data processing request; the target parameter information is identified to obtain the explanatory text corresponding to the data processing result; the explanatory text and the data processing result are fed back to the target object, which solves the problem that the AI ​​decision system in the related technology is a black box operation when processing the data processing request, resulting in poor interpretability of the data processing result. In summary, in this scheme, according to the data processing result after the artificial intelligence decision system processes the data processing request triggered by the target object and the target parameter information called when processing the target data processing request, an explanatory text corresponding to the data processing result is generated, and the explanatory text and the data processing result are synchronously fed back to the target object. Through the data processing results and the target parameter information, a complete track of the decision creation of the artificial intelligence decision system is created, so that the decision process of the artificial intelligence decision system is traceable and reviewable. At the same time, the explanatory text corresponding to the data processing result is fed back to the target object to help the target object understand the decision reasons of the artificial intelligence decision system and enhance the transparency of the artificial intelligence decision system.

[0068] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0069] The embodiment of the present application also provides a data processing device. It should be noted that the data processing device of the embodiment of the present application can be used to execute the data processing method provided by the embodiment of the present application. The data processing device provided by the embodiment of the present application is introduced below.

[0070] Figure 3 is a schematic diagram of a data processing device according to an embodiment of the present application. Figure 3 As shown, the device includes: an acquisition unit 301, an identification unit 302, and a feedback unit 303.

[0071] The acquisition unit 301 is used to obtain the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision-making system on the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision-making system when processing the target data processing request;

[0072] The identification unit 302 is used to identify the target parameter information and obtain an explanation text corresponding to the data processing result;

[0073] The feedback unit 303 is used to feed back the interpretation text and data processing results to the target object.

[0074] The data processing device provided by the embodiment of the present application, the acquisition unit 301 acquires the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision system when processing the target data processing request; the identification unit 302 identifies the target parameter information to obtain the explanatory text corresponding to the data processing result; the feedback unit 303 feeds back the explanatory text and the data processing result to the target object, which solves the problem that the AI ​​decision system in the related art is a black box operation when processing the data processing request, resulting in poor interpretability of the data processing result. In this scheme, according to the data processing result after the artificial intelligence decision system processes the data processing request triggered by the target object and the target parameter information called when processing the target data processing request, an explanatory text corresponding to the data processing result is generated, and the explanatory text and the data processing result are synchronously fed back to the target object. Through the data processing result and the target parameter information, a complete track of the decision created by the artificial intelligence decision system is created, so that the decision process of the artificial intelligence decision system is traceable and reviewable. At the same time, the explanatory text corresponding to the data processing result is fed back to the target object to help the target object understand the decision reason of the artificial intelligence decision system, and enhance the transparency of the artificial intelligence decision system.

[0075] Optionally, in the data processing device provided in the embodiment of the present application, the acquisition unit includes: a parsing module, which is used to parse and process the target data processing request through the data preprocessing unit of the artificial intelligence decision-making system to obtain target request data, wherein the target request data at least includes: the request time, request content and request context information of the target data processing request; a feature extraction module, which is used to extract features from the target request data through the feature extraction unit of the artificial intelligence decision-making system to obtain target feature data; and a first processing module, which is used to process the target data processing request based on the target feature data through the artificial intelligence decision-making system to obtain data processing results and target parameter information.

[0076] Optionally, in the data processing device provided in the embodiment of the present application, the recognition unit includes: a second processing module, used to perform word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; a third processing module, used to perform vectorization processing on the target parameter information after word segmentation to obtain the processed target parameter information; and a recognition module, used to recognize the processed target parameter information to obtain an explanation text.

[0077] Optionally, in the data processing device provided in the embodiment of the present application, the feedback unit includes: a judgment module, used to judge whether the interpretation text conforms to the preset execution logic, and whether the data processing result conforms to the expected result, to obtain a judgment result; a fourth processing module, used to visualize the interpretation text and the data processing result if the judgment result indicates that the interpretation text conforms to the preset execution logic and the data processing result conforms to the expected result, so as to feed back the interpretation text and the data processing result to the target object.

[0078] Optionally, in the data processing device provided in the embodiment of the present application, the device also includes: a first determination unit, for determining whether the interpretation text conforms to the preset execution logic and whether the data processing result conforms to the expected result, after obtaining the judgment result, if the judgment result indicates that the interpretation text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, then determining the deviation level corresponding to the interpretation text and the data processing result; a second determination unit, for determining a target adjustment strategy based on the deviation level; an adjustment unit, for adjusting the artificial intelligence decision system based on the target adjustment strategy, and obtaining the data processing result and the interpretation text again based on the adjusted artificial intelligence decision system.

[0079] Furthermore, the parsing module includes: a parsing submodule, which is used to parse the target data processing request to obtain initial request data; a cleaning submodule, which is used to clean the initial request data to obtain cleaned initial request data; and a filling submodule, which is used to fill in the missing parts of the cleaned initial request data to obtain the target request data.

[0080] Furthermore, the feature extraction module includes: a feature extraction submodule, which is used to perform feature extraction on the target request data through the convolution layer in the feature extraction unit to obtain initial feature data; a dimension reduction submodule, which is used to perform dimension reduction processing on the initial feature data through the pooling layer in the feature extraction unit to obtain the initial feature data after dimension reduction; and an output submodule, which is used to output the initial feature data after dimension reduction through the output layer in the feature extraction unit to obtain target feature data.

[0081] The data processing device includes a processor and a memory. The acquisition unit 301, the identification unit 302, the feedback unit 303, etc. are all stored in the memory as program units. The processor executes the program units stored in the memory to achieve efficient data processing.

[0082] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and efficient data processing can be achieved by adjusting kernel parameters.

[0083] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0084] An embodiment of the present invention provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, a data processing method is implemented.

[0085] An embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes a data processing method when it is run.

[0086] like Figure 4 As shown, an embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining a data processing result and target parameter information of a target data processing request triggered by an artificial intelligence decision-making system for a target object, wherein the target parameter information is parameter information called by the artificial intelligence decision-making system when processing the target data processing request; identifying the target parameter information to obtain an explanation text corresponding to the data processing result; and feeding back the explanation text and the data processing result to the target object.

[0087] Furthermore, obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object includes: parsing and processing the target data processing request through the data preprocessing unit of the artificial intelligence decision system to obtain target request data, wherein the target request data at least includes: the request time, request content and request context information of the target data processing request; extracting features from the target request data through the feature extraction unit of the artificial intelligence decision system to obtain target feature data; processing the target data processing request based on the target feature data through the artificial intelligence decision system to obtain data processing results and target parameter information.

[0088] Furthermore, the target parameter information is identified to obtain an explanatory text corresponding to the data processing result, including: performing word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; performing vectorization processing on the target parameter information after word segmentation to obtain the processed target parameter information; and identifying the processed target parameter information to obtain an explanatory text.

[0089] Furthermore, feeding back the interpretation text and the data processing results to the target object includes: judging whether the interpretation text complies with the preset execution logic, and whether the data processing results comply with the expected results, and obtaining a judgment result; if the judgment result indicates that the interpretation text complies with the preset execution logic, and the data processing results comply with the expected results, then visualizing the interpretation text and the data processing results to feed back the interpretation text and the data processing results to the target object.

[0090] Furthermore, after determining whether the explanation text conforms to the preset execution logic and whether the data processing result conforms to the expected result and obtaining the judgment result, the method also includes: if the judgment result indicates that the explanation text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, determining the deviation level corresponding to the explanation text and the data processing result; determining the target adjustment strategy based on the deviation level; adjusting the artificial intelligence decision system based on the target adjustment strategy, and obtaining the data processing result and the explanation text again based on the adjusted artificial intelligence decision system.

[0091] Furthermore, the target data processing request is parsed and processed by the data preprocessing unit of the artificial intelligence decision-making system to obtain the target request data, including: parsing the target data processing request to obtain initial request data; cleaning the initial request data to obtain cleaned initial request data; and filling in the missing parts of the cleaned initial request data to obtain the target request data.

[0092] Furthermore, the feature extraction unit of the artificial intelligence decision-making system performs feature extraction on the target request data to obtain the target feature data, including: performing feature extraction on the target request data through the convolution layer in the feature extraction unit to obtain initial feature data; performing dimensionality reduction processing on the initial feature data through the pooling layer in the feature extraction unit to obtain initial feature data after dimensionality reduction; and outputting the initial feature data after dimensionality reduction through the output layer in the feature extraction unit to obtain target feature data.

[0093] The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0094] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: obtaining the data processing results and target parameter information of a target data processing request triggered by an artificial intelligence decision-making system to a target object, wherein the target parameter information is the parameter information called by the artificial intelligence decision-making system when processing the target data processing request; identifying the target parameter information to obtain an explanatory text corresponding to the data processing result; and feeding back the explanatory text and the data processing result to the target object.

[0095] Furthermore, obtaining the data processing result and target parameter information of the target data processing request triggered by the artificial intelligence decision system to the target object includes: parsing and processing the target data processing request through the data preprocessing unit of the artificial intelligence decision system to obtain target request data, wherein the target request data at least includes: the request time, request content and request context information of the target data processing request; extracting features from the target request data through the feature extraction unit of the artificial intelligence decision system to obtain target feature data; processing the target data processing request based on the target feature data through the artificial intelligence decision system to obtain data processing results and target parameter information.

[0096] Furthermore, the target parameter information is identified to obtain an explanatory text corresponding to the data processing result, including: performing word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; performing vectorization processing on the target parameter information after word segmentation to obtain the processed target parameter information; and identifying the processed target parameter information to obtain an explanatory text.

[0097] Furthermore, feeding back the interpretation text and the data processing results to the target object includes: judging whether the interpretation text complies with the preset execution logic, and whether the data processing results comply with the expected results, and obtaining a judgment result; if the judgment result indicates that the interpretation text complies with the preset execution logic, and the data processing results comply with the expected results, then visualizing the interpretation text and the data processing results to feed back the interpretation text and the data processing results to the target object.

[0098] Furthermore, after determining whether the explanation text conforms to the preset execution logic and whether the data processing result conforms to the expected result and obtaining the judgment result, the method also includes: if the judgment result indicates that the explanation text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, determining the deviation level corresponding to the explanation text and the data processing result; determining the target adjustment strategy based on the deviation level; adjusting the artificial intelligence decision system based on the target adjustment strategy, and obtaining the data processing result and the explanation text again based on the adjusted artificial intelligence decision system.

[0099] Furthermore, the target data processing request is parsed and processed by the data preprocessing unit of the artificial intelligence decision-making system to obtain the target request data, including: parsing the target data processing request to obtain initial request data; cleaning the initial request data to obtain cleaned initial request data; and filling in the missing parts of the cleaned initial request data to obtain the target request data.

[0100] Furthermore, the feature extraction unit of the artificial intelligence decision-making system performs feature extraction on the target request data to obtain the target feature data, including: performing feature extraction on the target request data through the convolution layer in the feature extraction unit to obtain initial feature data; performing dimensionality reduction processing on the initial feature data through the pooling layer in the feature extraction unit to obtain initial feature data after dimensionality reduction; and outputting the initial feature data after dimensionality reduction through the output layer in the feature extraction unit to obtain target feature data.

[0101] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0102] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0103] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0104] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0105] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0106] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0107] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0108] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0109] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0110] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A data processing method, characterized in that: include: Obtaining data processing results and target parameter information of a target data processing request triggered by the artificial intelligence decision-making system on a target object, wherein the target parameter information is parameter information called by the artificial intelligence decision-making system when processing the target data processing request; Identify the target parameter information to obtain an explanation text corresponding to the data processing result; The interpretation text and the data processing result are fed back to the target object.

2. The method according to claim 1, characterized in that Obtaining the data processing results and target parameter information of the target data processing request triggered by the artificial intelligence decision-making system on the target object includes: The target data processing request is parsed and processed by the data preprocessing unit of the artificial intelligence decision-making system to obtain target request data, wherein the target request data at least includes: request time, request content and request context information of the target data processing request; Extracting features from the target request data by using a feature extraction unit of the artificial intelligence decision-making system to obtain target feature data; The target data processing request is processed by the artificial intelligence decision system based on the target feature data to obtain the data processing result and the target parameter information.

3. The method according to claim 1, characterized in that The target parameter information is identified, and the explanation text corresponding to the data processing result is obtained, including: Performing word segmentation processing on the target parameter information to obtain the target parameter information after word segmentation; Performing vectorization processing on the target parameter information after word segmentation to obtain processed target parameter information; The processed target parameter information is identified to obtain the explanation text.

4. The method according to claim 1, characterized in that: Feeding back the interpretation text and the data processing result to the target object includes: Determine whether the interpretation text conforms to the preset execution logic, and whether the data processing result conforms to the expected result, and obtain a determination result; If the judgment result indicates that the interpretation text complies with the preset execution logic and the data processing result complies with the expected result, the interpretation text and the data processing result are visualized to feed back the interpretation text and the data processing result to the target object.

5. The method according to claim 4, characterized in that After determining whether the interpretation text complies with the preset execution logic and whether the data processing result complies with the expected result, and obtaining the determination result, the method further includes: If the judgment result indicates that the interpretation text does not conform to the preset execution logic, or the data processing result does not conform to the expected result, then determining the deviation level corresponding to the interpretation text and the data processing result; Determine the target adjustment strategy based on the deviation level; The artificial intelligence decision system is adjusted according to the target adjustment strategy, and the data processing result and the explanation text are obtained again according to the adjusted artificial intelligence decision system.

6. The method according to claim 2, characterized in that The target data processing request is parsed and processed by the data preprocessing unit of the artificial intelligence decision-making system to obtain the target request data including: Parsing the target data processing request to obtain initial request data; Cleaning the initial request data to obtain cleaned initial request data; Perform missing filling processing on the cleaned initial request data to obtain the target request data.

7. The method according to claim 2, characterized in that The feature extraction unit of the artificial intelligence decision-making system extracts features from the target request data to obtain target feature data including: Extracting features of the target request data through a convolutional layer in the feature extraction unit to obtain initial feature data; Performing dimensionality reduction processing on the initial feature data through the pooling layer in the feature extraction unit to obtain initial feature data after dimensionality reduction; The reduced initial feature data is outputted through the output layer in the feature extraction unit to obtain the target feature data.

8. A data processing device, characterized in that: include: An acquisition unit, used to acquire a data processing result and target parameter information of a target data processing request triggered by the artificial intelligence decision-making system on a target object, wherein the target parameter information is parameter information called by the artificial intelligence decision-making system when processing the target data processing request; An identification unit, used to identify the target parameter information and obtain an explanation text corresponding to the data processing result; A feedback unit is used to feed back the interpretation text and the data processing result to the target object.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed, the storage medium is controlled to execute the data processing method according to any one of claims 1 to 7 on a device.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the data processing method described in any one of claims 1 to 7.

Images