Model optimization method and device, equipment, medium and program product

Abstract

The invention provides a model optimization method and device, equipment, a medium and a program product, which can be applied to the field of artificial intelligence and the field of finance. The method comprises the following steps: acquiring basic service data corresponding to a plurality of different service scenes; generating a rewriting rule aiming at the basic business data in each business scene; based on a rewriting rule, a natural language processing model is utilized to rewrite the basic business data, target business data suitable for all the business scenes are generated, and semantics of the target business data and semantics of the basic business data are consistent; evaluating the to-be-optimized model by using the target business data to obtain an evaluation result; and according to the evaluation result, optimizing the to-be-optimized model to obtain a target optimization model.

Description

Technical Field

[0001] The present disclosure relates to the fields of artificial intelligence and finance, and more specifically to a model optimization method, device, equipment, medium and program product. Background Art

[0002] In the financial field, large-scale machine learning models are often used for risk prediction, business management and other operations. Among them, model evaluation and optimization are crucial in the process of model application.

[0003] In related technologies, fixed data set evaluation methods and resampling techniques are often used to evaluate models, so as to optimize models based on the evaluation results. However, using a preset fixed data set to test models has the risk of selection bias and is difficult to adapt to a dynamically changing environment, especially in the rapidly changing financial field. It often shows poor adaptability, resulting in low accuracy of evaluation results, and thus poor model optimization effects. Summary of the invention

[0004] In view of the above problems, the present disclosure provides a model optimization method, apparatus, device, medium and program product.

[0005] According to a first aspect of the present disclosure, a model optimization method is provided, including: acquiring basic business data corresponding to a plurality of different business scenarios; generating rewriting rules for the basic business data in each business scenario; based on the rewriting rules, rewriting the basic business data using a natural language processing model to generate target business data applicable to each of the business scenarios, wherein the semantics of the target business data and the basic business data are consistent; evaluating the model to be optimized using the target business data to obtain an evaluation result; optimizing the model to be optimized based on the evaluation result to obtain a target optimization model.

[0006] According to an embodiment of the present disclosure, the rewriting rules include replacement sub-rules, structure sub-rules and scenario sub-rules; wherein the replacement sub-rules are used to determine the synonyms of the terms of the basic business data; the structure sub-rules are used to determine the adjustment method of the structural order of the basic business data; and the scenario sub-rules are used to determine the scenario terms corresponding to the basic business data.

[0007] According to an embodiment of the present disclosure, based on the rewriting rules, the basic business data is rewritten using a natural language processing model to generate target business data, including: based on the replacement sub-rule, the terms of the basic business data are replaced with synonyms using a natural language processing model to generate first business data; based on the structure sub-rule, the structural order of the first business data is adjusted using a natural language processing model to generate second business data; based on the scenario sub-rule, the scenario terms of the second business data are changed using a natural language processing model to produce target business data.

[0008] According to an embodiment of the present disclosure, the above method also includes: optimizing the rewriting rules according to the evaluation results to obtain optimized rewriting rules; based on the optimized rewriting rules, optimizing and rewriting the basic business data using a natural language processing model.

[0009] According to an embodiment of the present disclosure, the rewriting rules are optimized according to the evaluation results, including: determining the optimization target that the evaluation results are intended to meet according to the evaluation results; based on the optimization target, optimizing the rewriting rules using an optimization algorithm to obtain optimized rewriting rules.

[0010] According to an embodiment of the present disclosure, the model to be optimized is evaluated using target business data to obtain evaluation results, including: determining a target evaluation operation corresponding to the business scenario based on the business scenario of the model to be optimized; and evaluating the model to be optimized based on the target evaluation operation to obtain evaluation results.

[0011] According to an embodiment of the present disclosure, the above method also includes: using a task scheduling tool to determine the update cycle; based on the update cycle, using a natural language processing model to update and rewrite the basic business data to obtain updated target business data; based on the updated target business data, updating and evaluating the model to be optimized to obtain updated evaluation results.

[0012] According to an embodiment of the present disclosure, the model to be optimized is optimized according to the evaluation results to obtain a target optimization model, including: quantitatively analyzing the evaluation results to obtain analysis results; optimizing the model to be optimized according to the analysis results to obtain a target optimization model.

[0013] The second aspect of the present disclosure provides a model optimization device, including: an acquisition module, used to acquire basic business data corresponding to multiple different business scenarios; a first generation module, used to generate rewriting rules for the basic business data in each business scenario; a second generation module, used to rewrite the basic business data based on the rewriting rules using a natural language processing model to generate target business data, wherein the target business data and the basic business data have consistent semantics; an evaluation module, used to evaluate the model to be optimized using the target business data to obtain an evaluation result; and a first optimization module, used to optimize the model to be optimized according to the evaluation result to obtain a target optimization model.

[0014] A third aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more computer programs, wherein the one or more processors execute the one or more computer programs to implement the steps of the above method.

[0015] The fourth aspect of the present disclosure further provides a computer-readable storage medium having a computer program or instructions stored thereon, which implements the steps of the above method when the above computer program or instructions are executed by a processor.

[0016] The fifth aspect of the present disclosure further provides a computer program product, including a computer program or instructions, which implement the steps of the above method when the above computer program or instructions are executed by a processor.

[0017] According to the embodiments of the present disclosure, basic business data corresponding to multiple different business scenarios are obtained; rewriting rules for the basic business data in each business scenario are generated; based on the rewriting rules, the basic business data are rewritten using a natural language processing model to generate target business data, wherein the semantics of the target business data and the basic business data are consistent; the target business data is used to evaluate the model to be optimized to obtain the evaluation results; based on the evaluation results, the model to be optimized is optimized to obtain the target optimization model. Since specific rewriting rules are generated for different business scenarios, the target business data can be generated using a natural language processing model without changing the semantics, thus avoiding the problem in the prior art that the data is fixed and difficult to adapt to the scenarios in the financial field, achieving high accuracy of the evaluation results, avoiding model evaluation deviations, and thus improving the technical effect of the optimization effect of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0019] Figure 1 The application scenario diagram of the model optimization method, apparatus, device, medium and program product according to the embodiment of the present disclosure is schematically shown;

[0020] Figure 2 A flow chart of a model optimization method according to an embodiment of the present disclosure is schematically shown;

[0021] Figure 3 A schematic diagram schematically shows a model optimization method according to an embodiment of the present disclosure;

[0022] Figure 4 A structural block diagram of a model optimization device according to an embodiment of the present disclosure is schematically shown; and

[0023] Figure 5 A block diagram of an electronic device suitable for implementing a model optimization method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0024] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0025] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0026] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0027] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0028] In the technical solution of the present disclosure, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0029] In the scenario of using personal information for automated decision-making, the methods, devices, and systems provided by the embodiments of the present disclosure provide users with corresponding operation portals for users to choose to agree or reject the automated decision-making results; if the user chooses to reject, the expert decision-making process will be entered. The expression "automated decision-making" here refers to the activity of automatically analyzing and evaluating a person's behavioral habits, interests and hobbies, or economic, health, credit status, etc. through computer programs, and making decisions. The expression "expert decision-making" here refers to the activity of making decisions by people who specialize in a certain field, have specialized experience, knowledge and skills, and have reached a certain level of professionalism.

[0030] The embodiments of the present disclosure provide a model optimization method, which obtains basic business data corresponding to multiple different business scenarios; generates rewriting rules for the basic business data in each business scenario; based on the rewriting rules, rewrites the basic business data using a natural language processing model to generate target business data, wherein the target business data and the basic business data have the same semantics; evaluates the model to be optimized using the target business data to obtain evaluation results; and optimizes the model to be optimized based on the evaluation results to obtain a target optimization model.

[0031] Figure 1 The application scenario diagram of the model optimization method, apparatus, device, medium and program product according to the embodiments of the present disclosure is schematically shown.

[0032] like Figure 1 As shown, the application scenario 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0033] The user can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only for example).

[0034] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0035] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0036] It should be noted that the model optimization method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the model optimization device provided in the embodiment of the present disclosure can generally be set in the server 105. The model optimization method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Accordingly, the model optimization device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105.

[0037] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0038] The following will be based on Figure 1 The described scenario describes the model optimization method of the disclosed embodiment in detail.

[0039] Figure 2 The flowchart of the model optimization method according to the embodiment of the present disclosure is schematically shown.

[0040] like Figure 2 As shown, the model optimization method of this embodiment includes operations S210 to S250.

[0041] In operation S210, basic business data corresponding to a plurality of different business scenarios is obtained.

[0042] According to the embodiments of the present disclosure, basic business data can be financial statements for different businesses in the same business scenario, for example, statements such as "the company's profit increased by 20% in the second quarter of 2023" used to input models to implement risk prediction, information management and other operations based on the model, wherein the source of basic business data can be market data, bank transaction records, debt and asset information, company financial reports, etc. To improve the timeliness of basic business data, a real-time data capture system can be combined to dynamically collect the latest changes in the financial market. In addition, by acquiring real-time financial news, news fact data can be introduced into basic business data to increase the responsiveness of basic business data to the latest financial events, ensuring that basic business data fully and timely reflects the actual situation of the current market, thereby improving the authenticity of the evaluation of the model to be optimized and the relevance of the application.

[0043] According to an embodiment of the present disclosure, the business scenario may be a banking service scenario such as “amount transaction” or “securities business”.

[0044] In operation S220, rewriting rules for basic business data in each business scenario are generated.

[0045] According to an embodiment of the present disclosure, the rewriting rules are used to ensure that the semantics of basic business data remain unchanged during the rewriting process.

[0046] In operation S230 , the basic business data is rewritten based on the rewriting rule using a natural language processing model to generate target business data, wherein the target business data and the basic business data have the same semantics.

[0047] According to an embodiment of the present disclosure, the natural language processing model may be a deep learning model architecture.

[0048] According to the embodiments of the present disclosure, the natural language processing model can extract relevant financial events, market sentiment and other information from financial news. These data can be dynamically added to the existing data set, so that the generated data set is closer to the current market status, thereby improving the real-time and representativeness of the evaluation data.

[0049] According to the embodiments of the present disclosure, the basic business data is rewritten based on the rewriting rules using a natural language processing model, so that the integrity of the original financial information can be maintained when the text is rewritten. Especially in scenarios involving complex financial logic, the rewritten text is guaranteed to be highly semantically consistent with the original text, thereby generating target business data.

[0050] In operation S240, the model to be optimized is evaluated using the target business data to obtain an evaluation result.

[0051] According to an embodiment of the present disclosure, for models to be optimized in different financial scenarios, target business data is used to evaluate the models to be optimized to obtain evaluation results suitable for different financial scenarios.

[0052] In operation S250, the model to be optimized is optimized according to the evaluation result to obtain a target optimization model.

[0053] According to an embodiment of the present disclosure, an optimization target associated with the model to be optimized is set based on the evaluation result to optimize the model to be optimized so that the obtained target optimization model can meet the optimization target.

[0054] According to the embodiments of the present disclosure, basic business data corresponding to multiple different business scenarios are obtained; rewriting rules for the basic business data in each business scenario are generated; based on the rewriting rules, the basic business data are rewritten using a natural language processing model to generate target business data, wherein the semantics of the target business data and the basic business data are consistent; the target business data is used to evaluate the model to be optimized to obtain the evaluation results; based on the evaluation results, the model to be optimized is optimized to obtain the target optimization model. Since specific rewriting rules are generated for different business scenarios, the target business data can be generated using a natural language processing model without changing the semantics, thus avoiding the problem in the prior art that the data is fixed and difficult to adapt to the scenarios in the financial field, achieving high accuracy of the evaluation results, avoiding model evaluation deviations, and thus improving the technical effect of the optimization effect of the model.

[0055] According to an embodiment of the present disclosure, the rewriting rules include replacement sub-rules, structure sub-rules and scenario sub-rules; wherein the replacement sub-rules are used to determine the synonyms of the terms of the basic business data; the structure sub-rules are used to determine the adjustment method of the structural order of the basic business data; and the scenario sub-rules are used to determine the scenario terms corresponding to the basic business data.

[0056] According to an embodiment of the present disclosure, the replacement sub-rule is used to replace the financial terms and common words in the basic business data with synonyms to enrich the expressions while maintaining semantic consistency. For example, "return on investment" is replaced with "yield".

[0057] According to an embodiment of the present disclosure, the structural sub-rule is used to adjust the grammatical structure or sentence order of the basic business data to increase the complexity and diversity of the data. For example, "the company's profit increased by 20% in the second quarter of 2023" can be rewritten as "the company achieved a 20% profit growth in the second quarter of 2023".

[0058] According to an embodiment of the present disclosure, scenario sub-rules are used to appropriately adjust the use of financial terms in basic business data for different business application scenarios to better conform to the financial background of specific business application scenarios. For example, "debt repayment capacity" can be replaced with "debt repayment capacity" to ensure the accuracy of terminology used in different financial contexts. For another example, the design of "exchange rate" in the basic business data involves "exchange rate" for the business scenario of "only one exchange rate for a pair of foreign countries", and "exchange rate" is changed to "single rate (Single Rate)", and for the business scenario of "multiple exchange rates for a pair of foreign countries", "exchange rate" is changed to "multiple rate (Multiple Rate)".

[0059] According to an embodiment of the present disclosure, based on the rewriting rules, the basic business data is rewritten using a natural language processing model to generate target business data, including: based on the replacement sub-rule, the terms of the basic business data are replaced with synonyms using a natural language processing model to generate first business data; based on the structure sub-rule, the structural order of the first business data is adjusted using a natural language processing model to generate second business data; based on the scenario sub-rule, the scenario terms of the second business data are changed using a natural language processing model to generate target business data.

[0060] According to an embodiment of the present disclosure, based on the above-mentioned replacement sub-rules, replacement sub-rules and scenario sub-rules, the basic business data is rewritten using a natural language processing model, which can maintain the integrity of the basic business data, especially in scenarios involving complex financial logic, to ensure that the rewritten text is highly semantically consistent with the original text.

[0061] According to the embodiments of the present disclosure, after the basic business data is rewritten using a natural language processing model, in order to ensure that the rewritten or dynamically generated target business data maintains its compliance and validity, it is necessary to perform strict compliance verification on the target business data after it is generated to ensure that the target business data complies with actual financial standards and industry specifications and can reflect the dynamic environment of the current financial market.

[0062] According to an embodiment of the present disclosure, the rewriting order of the above-mentioned replacement sub-rules, substitution sub-rules and scenario sub-rules can be randomly set.

[0063] According to the embodiments of the present disclosure, by rewriting the basic business data based on the rewriting rules and using the natural language processing model to dynamically generate the target business data, the security and confidentiality of the data can be significantly improved. In addition, the generated data sets are all unique, thereby avoiding the risk of leakage caused by fixed data sets and improving the security of model evaluation and optimization. At the same time, the generation mechanism of the natural language processing model ensures the freshness and diversity of the target business data, so that the generated target business data can more truly reflect the changes in the financial market, improve the accuracy of the evaluation results, and avoid the problem of model evaluation bias.

[0064] According to an embodiment of the present disclosure, the above method also includes: optimizing the rewriting rules according to the evaluation results to obtain optimized rewriting rules; based on the optimized rewriting rules, optimizing and rewriting the basic business data using a natural language processing model.

[0065] According to an embodiment of the present disclosure, when the evaluation results are abnormal, such as when they differ greatly from historical evaluation results, it is determined that the current rewriting rules are not applicable to the model to be optimized, and the rewriting rules need to be optimized to update the target business data.

[0066] According to an embodiment of the present disclosure, the rewriting rules are optimized according to the evaluation results, including: determining the optimization target that the evaluation results are intended to meet according to the evaluation results; based on the optimization target, optimizing the rewriting rules using an optimization algorithm to obtain optimized rewriting rules.

[0067] According to the embodiments of the present disclosure, the optimization target may be determined based on historical evaluation results, and may also be obtained by further adjusting the evaluation results by those skilled in the art.

[0068] According to an embodiment of the present disclosure, the rewriting rules are optimized using an optimization algorithm to obtain new rewriting rules, so that the target business data generated based on the new rewriting rules and the evaluation results obtained by evaluating the model to be optimized meet the optimization goals.

[0069] According to an embodiment of the present disclosure, the model to be optimized is evaluated using target business data to obtain evaluation results, including: determining a target evaluation operation corresponding to the business scenario based on the business scenario of the model to be optimized; and evaluating the model to be optimized based on the target evaluation operation to obtain evaluation results.

[0070] According to an embodiment of the present disclosure, in order to measure the performance of the model to be optimized in the prediction classification task, the "Accuracy" target evaluation operation is used to evaluate the model to be optimized to obtain an evaluation result.

[0071] According to the embodiments of the present disclosure, for business scenarios such as financial risk prediction, the "F1 score" target evaluation operation is used to evaluate the model to be optimized to obtain the evaluation results.

[0072] According to an embodiment of the present disclosure, for business scenarios such as predicting market prices or financial indicators, the "mean square error (MSE)" target evaluation operation is used to evaluate the model to be optimized to obtain an evaluation result.

[0073] According to the embodiments of the present disclosure, for business scenarios such as binary classification problems such as credit rating, the "ROC curve" target evaluation operation is used to evaluate the model to be optimized to obtain an evaluation result.

[0074] According to the embodiments of the present disclosure, a comprehensive performance evaluation is performed on the model to be optimized. By testing in a variety of business scenarios, it is ensured that the model to be optimized can run stably and effectively under different market conditions to verify its generalization ability and practicality. The evaluation content not only includes the evaluation of prediction accuracy, but also focuses on the adaptability and robustness of the model to be optimized in different financial scenarios to improve the accuracy of the evaluation results of the model to be optimized in different business scenarios.

[0075] According to an embodiment of the present disclosure, the above method also includes: using a task scheduling tool to determine the update cycle; based on the update cycle, using a natural language processing model to update and rewrite the basic business data to obtain updated target business data; based on the updated target business data, updating and evaluating the model to be optimized to obtain updated evaluation results.

[0076] According to the embodiments of the present disclosure, a task scheduling tool is used to achieve regular generation and update of target business data, and an intelligent generation strategy is used to reduce manual intervention, improve the degree of automation of the generation process, and ensure the diversity and timeliness of the target business data.

[0077] According to the embodiments of the present disclosure, the rewriting rules are regularly optimized in combination with market feedback and changes in economic conditions to ensure that the target business data can still reflect key features when facing drastic market fluctuations (such as economic crises, etc.), such as adding data related to market volatility and risk management during economic crises.

[0078] According to the embodiments of the present disclosure, through the automated data generation and optimization mechanism, the evaluation process is made more efficient and flexible, and can respond to market changes in a timely manner, ensuring that the evaluation of the model to be optimized keeps pace with the times.

[0079] According to an embodiment of the present disclosure, the model to be optimized is optimized according to the evaluation results to obtain a target optimization model, including: quantitatively analyzing the evaluation results to obtain analysis results; optimizing the model to be optimized according to the analysis results to obtain a target optimization model.

[0080] According to an embodiment of the present disclosure, quantitative analysis is a method of analyzing financial data such as financial markets and securities using mathematical models, statistical methods and algorithms, and involves using historical evaluation results to predict future market behavior to obtain analysis results.

[0081] According to the embodiments of the present disclosure, after the evaluation is completed, the evaluation results are quantitatively analyzed and fed back to the model to be optimized. The data generation strategy and model parameters are adjusted based on the performance differences found in the evaluation. This feedback mechanism ensures that the generated data can be continuously improved, thereby driving the model to achieve better performance in future tests.

[0082] Figure 3 A schematic diagram of a model optimization method according to an embodiment of the present disclosure is schematically shown.

[0083] According to an embodiment of the present disclosure, the model optimization method of the embodiment of the present disclosure is as follows: Figure 3 As shown, perform operation S310 to obtain basic business data; then perform operation S320 to generate rewriting rules corresponding to the basic business data; perform operation S330 to rewrite the basic business data using a natural language processing model based on the rewriting rules generated above to generate target business data; perform operation S340 to determine a target evaluation operation based on the business scenario of the model to be optimized, and then perform operation S350 to evaluate the model to be optimized based on the target evaluation operation to obtain an evaluation result; perform operation S360 to determine an optimization target based on the evaluation result; perform operation S370 to optimize the rewriting rules, and the optimized rewriting rules obtained make the evaluation result of the model to be optimized meet the optimization target; perform operation S380 to optimize the model to be optimized based on the evaluation result to obtain a target optimization model.

[0084] Based on the above model optimization method, the present disclosure also provides a model optimization device. Figure 4 The device is described in detail.

[0085] Figure 4 The structural block diagram of the model optimization device according to an embodiment of the present disclosure is schematically shown.

[0086] like Figure 4 As shown, the model optimization device 400 of this embodiment includes an acquisition module 410, a first generation module 420, a second generation module 430, an evaluation module 440 and a first optimization module 450.

[0087] The acquisition module 410 is used to acquire basic business data corresponding to a plurality of different business scenarios. In one embodiment, the acquisition module 410 can be used to perform the operation S210 described above, which will not be described in detail here.

[0088] The first generating module 420 is used to generate rewriting rules for basic business data in various business scenarios. In one embodiment, the first generating module 420 can be used to perform the operation S220 described above, which will not be described in detail here.

[0089] The second generation module 430 is used to rewrite the basic business data based on the rewriting rule using the natural language processing model to generate target business data, wherein the target business data and the basic business data have the same semantics. In one embodiment, the second generation module 430 can be used to perform the operation S230 described above, which will not be repeated here.

[0090] The evaluation module 440 is used to evaluate the model to be optimized using the target business data to obtain an evaluation result. In one embodiment, the evaluation module 440 can be used to perform the operation S240 described above, which will not be described in detail here.

[0091] The first optimization module 450 is used to optimize the model to be optimized according to the evaluation result to obtain the target optimization model. In one embodiment, the first optimization module 450 can be used to perform the operation S250 described above, which will not be described in detail here.

[0092] According to the embodiments of the present disclosure, basic business data corresponding to multiple different business scenarios are obtained; rewriting rules for the basic business data in each business scenario are generated; based on the rewriting rules, the basic business data are rewritten using a natural language processing model to generate target business data, wherein the semantics of the target business data and the basic business data are consistent; the target business data is used to evaluate the model to be optimized to obtain the evaluation results; based on the evaluation results, the model to be optimized is optimized to obtain the target optimization model. Since specific rewriting rules are generated for different business scenarios, the target business data can be generated using a natural language processing model without changing the semantics, thus avoiding the problem in the prior art that the data is fixed and difficult to adapt to the scenarios in the financial field, achieving high accuracy of the evaluation results, avoiding model evaluation deviations, and thus improving the technical effect of the optimization effect of the model.

[0093] According to an embodiment of the present disclosure, any multiple modules among the acquisition module 410, the first generation module 420, the second generation module 430, the evaluation module 440 and the first optimization module 450 can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the acquisition module 410, the first generation module 420, the second generation module 430, the evaluation module 440 and the first optimization module 450 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the acquisition module 410, the first generation module 420, the second generation module 430, the evaluation module 440 and the first optimization module 450 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0094] According to an embodiment of the present disclosure, the second generation module 430 includes a first generation submodule, a second generation submodule and a third generation submodule.

[0095] The first generating submodule is used to replace the terms of the basic business data with synonyms based on the replacement sub-rule by using a natural language processing model to generate the first business data.

[0096] The second generating submodule is used to adjust the structural sequence of the first business data based on the structural sub-rule and using the natural language processing model to generate the second business data.

[0097] The third generating submodule is used to modify the scenario terminology of the second business data based on the scenario sub-rule and using the natural language processing model to generate target business data.

[0098] According to an embodiment of the present disclosure, the model optimization device 400 further includes a second optimization module and a rewriting module.

[0099] The second optimization module is used to optimize the rewriting rules according to the evaluation results to obtain optimized rewriting rules.

[0100] The rewriting module is used to optimize and rewrite basic business data based on the optimized rewriting rules using a natural language processing model.

[0101] According to an embodiment of the present disclosure, the rewriting module includes a first determining submodule and a first optimizing submodule.

[0102] The first determination submodule is used to determine the optimization target that the evaluation result is intended to meet based on the evaluation result.

[0103] The first optimization submodule is used to optimize the rewriting rules based on the optimization target by using the optimization algorithm to obtain the optimized rewriting rules.

[0104] According to an embodiment of the present disclosure, the evaluation module 440 includes a second determination submodule and an evaluation submodule.

[0105] The second determination submodule is used to determine the target evaluation operation corresponding to the business scenario according to the business scenario of the model to be optimized.

[0106] The evaluation submodule is used to evaluate the model to be optimized based on the target evaluation operation and obtain the evaluation results.

[0107] According to an embodiment of the present disclosure, the model optimization device 400 further includes a first determining module, a first obtaining module and a second obtaining module.

[0108] The first determination module is used to determine the update cycle by using a task scheduling tool.

[0109] The first obtaining module is used to update and rewrite the basic business data based on the update cycle using the natural language processing model to obtain the updated target business data.

[0110] The second obtaining module is used to update and evaluate the optimization model based on the updated target business data to obtain updated evaluation results.

[0111] According to an embodiment of the present disclosure, the first optimization module 450 includes an analysis submodule and a third obtaining submodule.

[0112] The analysis submodule is used to perform quantitative analysis on the evaluation results to obtain analysis results.

[0113] The third submodule is used to optimize the model to be optimized according to the analysis results to obtain the target optimization model.

[0114] Figure 5 A block diagram of an electronic device suitable for implementing a model optimization method according to an embodiment of the present disclosure is schematically shown.

[0115] like Figure 5As shown, the electronic device 500 according to an embodiment of the present disclosure includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 to a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include an onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0116] In RAM 903, various programs and data required for the operation of electronic device 500 are stored. Processor 501, ROM 502 and RAM 503 are connected to each other via bus 504. Processor 501 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 502 and / or RAM 503. It should be noted that the program can also be stored in one or more memories other than ROM 502 and RAM 503. Processor 501 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.

[0117] According to an embodiment of the present disclosure, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to the bus 504. The electronic device 500 may further include one or more of the following components connected to the input / output (I / O) interface 505: an input portion 506 including a keyboard, a mouse, etc.; an output portion 507 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 508 including a hard disk, etc.; and a communication portion 509 including a network interface card such as a LAN card, a modem, etc. The communication portion 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output (I / O) interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 510 as needed, so that a computer program read therefrom is installed into the storage portion 508 as needed.

[0118] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.

[0119] According to an embodiment of the present disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device. For example, according to an embodiment of the present disclosure, the computer-readable storage medium may include the ROM 402 and / or RAM 403 described above and / or one or more memories other than ROM 402 and RAM 403.

[0120] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the model optimization method provided by the embodiment of the present disclosure.

[0121] The above functions defined in the system / device of the embodiment of the present disclosure are performed when the computer program is executed by the processor 501. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0122] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 509, and / or installed from the removable medium 511. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0123] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 509, and / or installed from the removable medium 511. When the computer program is executed by the processor 501, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.

[0124] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0125] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0126] It will be appreciated by those skilled in the art that the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations or combinations are not explicitly described in the present disclosure. In particular, without departing from the spirit and teachings of the present disclosure, the features described in the various embodiments of the present disclosure may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present disclosure.

[0127] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A model optimization method, characterized in that: The method comprises: Obtain basic business data corresponding to multiple different business scenarios; Generating rewriting rules for the basic business data in each of the business scenarios; Based on the rewriting rules, the basic business data is rewritten using a natural language processing model to generate target business data applicable to each of the business scenarios, wherein the semantics of the target business data and the basic business data are consistent; Using the target business data to evaluate the model to be optimized to obtain an evaluation result; According to the evaluation results, the model to be optimized is optimized to obtain a target optimization model.

2. The method according to claim 1, characterized in that The rewriting rules include replacement sub-rules, structure sub-rules and scenario sub-rules; Wherein, the replacement sub-rule is used to determine the synonyms of the terms of the basic business data; The structural sub-rule is used to determine the adjustment method of the basic business data structure sequence; The scenario sub-rule is used to determine the scenario term corresponding to the basic business data.

3. The method according to claim 2, characterized in that The rewriting rules are based on which the basic business data is rewritten using a natural language processing model to generate target business data applicable to each of the business scenarios, including: Based on the replacement sub-rule, using the natural language processing model, synonym replacement is performed on the terms of the basic business data to generate first business data; Based on the structural sub-rule, the natural language processing model is used to adjust the structural order of the first business data to generate second business data; Based on the scenario sub-rule, the scenario terminology of the second business data is modified by using the natural language processing model to generate the target business data.

4. The method according to claim 1, characterized in that: The method further comprises: According to the evaluation results, the rewriting rules are optimized to obtain optimized rewriting rules; Based on the optimized rewriting rules, the basic business data is optimized and rewritten using a natural language processing model.

5. The method according to claim 4, characterized in that The step of optimizing the rewriting rules according to the evaluation results includes: According to the evaluation results, determining the optimization objectives that the evaluation results are intended to meet; Based on the optimization target, the rewriting rule is optimized using an optimization algorithm to obtain the optimized rewriting rule.

6. The method according to claim 1, characterized in that The using the target business data to evaluate the model to be optimized to obtain an evaluation result includes: According to the business scenario of the model to be optimized, determining the target evaluation operation corresponding to the business scenario; Based on the target evaluation operation, the model to be optimized is evaluated to obtain the evaluation result.

7. The method according to claim 1, characterized in that The method further comprises: Use task scheduling tools to determine update cycles; Based on the update cycle, the basic business data is updated and rewritten using a natural language processing model to obtain updated target business data; Based on the updated target business data, the model to be optimized is updated and evaluated to obtain an updated evaluation result.

8. The method according to claim 1, characterized in that The step of optimizing the model to be optimized according to the evaluation result to obtain a target optimization model includes: Performing quantitative analysis on the evaluation results to obtain analysis results; The model to be optimized is optimized according to the analysis result to obtain the target optimization model.

9. A model optimization device, characterized in that: The device comprises: An acquisition module is used to acquire basic business data corresponding to multiple different business scenarios; A first generating module, used to generate a rewriting rule for the basic business data in each of the business scenarios; A second generating module is used to rewrite the basic business data based on the rewriting rule by using a natural language processing model to generate target business data, wherein the target business data and the basic business data have the same semantics; An evaluation module, used to evaluate the model to be optimized using the target business data to obtain an evaluation result; and The first optimization module is used to optimize the model to be optimized according to the evaluation result to obtain a target optimization model.

10. An electronic device comprising: one or more processors; a memory for storing one or more computer programs, It is characterized in that the one or more processors execute the one or more computer programs to implement the steps of the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images