File compliance checking method and device, electronic equipment and storage medium

Abstract

The application relates to the technical field of artificial intelligence, and provides a file compliance checking method and device, electronic equipment and a storage medium, the method comprises the following steps: obtaining a compliance file to be checked; performing first preprocessing on the compliance file to be checked, and obtaining a compliance type; based on the compliance type, obtaining an original compliance file sample set; performing second preprocessing on the original compliance file sample set, and obtaining a standard compliance file sample set; creating a compliance checking model, inputting the standard compliance file sample set and the compliance file to be checked into the compliance checking model, and obtaining a compliance checking result, wherein the compliance checking model comprises a semantic model and a compliance component model. The standard compliance file is recombined with the standardized compliance rule output by the semantic model to train the compliance component model, so that the compliance component model obtained through training is more accurate, and the accuracy of the compliance checking result is improved.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and specifically to a document compliance inspection method, apparatus, electronic device, and storage medium. Background Technology

[0002] Whether existing compliance documents meet compliance requirements requires extensive manual sampling checks by the company itself or regulatory authorities.

[0003] However, with the continuous promulgation of existing compliance rules, the existing system and the new compliance rules need to go through multiple stages. During the inspection process, many rely on the user's experience and ability, which may not be able to conduct the inspection objectively, resulting in a low accuracy rate of compliance document inspection.

[0004] Therefore, it is necessary to propose a method for checking compliance documents quickly and accurately. Summary of the Invention

[0005] In view of the above, it is necessary to propose a document compliance inspection method, device, electronic device and storage medium to recombine standard compliance documents and standardized compliance rules output by semantic model to train the compliance component model, so that the trained compliance component model is more accurate, thereby improving the accuracy of compliance inspection results.

[0006] A first aspect of the present invention provides a document compliance check method, the method comprising:

[0007] Parse the received document compliance check request to obtain the compliance document to be checked;

[0008] The compliance document to be inspected undergoes a first preprocessing step to obtain the compliance type corresponding to the compliance document to be inspected;

[0009] Based on the compliance type corresponding to the compliance documents to be inspected, obtain the original compliance document sample set;

[0010] The original compliance document sample set is subjected to a second preprocessing step to obtain a standard compliance document sample set;

[0011] A compliance inspection model is created by inputting the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected. The compliance inspection model includes a semantic model and a compliance component model.

[0012] Optionally, the first preprocessing of the compliance document to be inspected to obtain the compliance type corresponding to the compliance document to be inspected includes:

[0013] Identify the tags on the compliance documents to be inspected;

[0014] The identified labels are input into a pre-trained compliance type classification model to obtain the compliance type corresponding to the compliance document to be inspected.

[0015] Optionally, obtaining the original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected includes:

[0016] Obtain the target data source corresponding to the compliance type of the compliance document to be inspected from the preset source database;

[0017] The web crawler technology is used to obtain a sample set of original compliant documents from the target data source according to preset rules.

[0018] Optionally, the second preprocessing of the original compliance document sample set to obtain the standard compliance document sample set includes:

[0019] Based on the compliance type of the compliance document to be inspected, obtain the corresponding standard-defined data structure;

[0020] According to the data structure, each original compliance document sample in the original compliance document sample set is transformed to obtain a standard compliance document sample set.

[0021] Optionally, the step of inputting the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected includes:

[0022] The standard compliance document sample set is input into the semantic model to obtain standardized compliance rules;

[0023] The standardized compliance rules and the compliance documents to be checked are input into the compliance component model connected to the semantic model to obtain the compliance check results of the compliance documents to be checked.

[0024] Optionally, the training process of the semantic model includes:

[0025] Obtain multiple historical standard compliance documents corresponding to the compliance type and a standardized compliance rule set for each historical standard compliance document;

[0026] Find the semantic vector corresponding to each standardized compliance rule in the standardized compliance rules of each historical standard compliance document from the preset semantic table, and construct the semantic matrix of each historical standard compliance document;

[0027] The semantic matrices of the aforementioned historical standard compliance documents are determined as a semantic sample set;

[0028] The semantic sample set is divided into a training set and a test set according to a preset ratio;

[0029] The training set is input into a preset neural network for training to obtain a semantic model;

[0030] The test set is input into the semantic model for testing, and the test pass rate is calculated.

[0031] If the test pass rate is greater than or equal to a preset pass rate threshold, the semantic model training is considered complete; if the test pass rate is less than the preset pass rate threshold, the number of training sets is increased, and the semantic model is retrained.

[0032] Optionally, the step of inputting the standardized compliance rules and the compliance document to be inspected into the compliance component model connected to the semantic model to obtain the compliance inspection result of the compliance document to be inspected includes:

[0033] The compliance document to be inspected is converted according to the data structure defined in the standard corresponding to the compliance type to obtain the standard compliance document to be inspected;

[0034] The standard compliance document to be checked and the standardized compliance rules are input into the compliance component model connected to the semantic model to obtain the compliance check result of the compliance document to be checked.

[0035] A second aspect of the present invention provides a document compliance checking apparatus, the apparatus comprising:

[0036] The parsing and retrieval module is used to parse the received file compliance check requests and retrieve the compliance files to be checked;

[0037] The first preprocessing module is used to perform a first preprocessing on the compliance document to be inspected and obtain the compliance type corresponding to the compliance document to be inspected.

[0038] The acquisition module is used to acquire an original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected;

[0039] The second preprocessing module is used to perform a second preprocessing on the original compliance document sample set to obtain a standard compliance document sample set.

[0040] A creation and input module is used to create a compliance inspection model. The standard compliance document sample set and the compliance document to be inspected are input into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected. The compliance inspection model includes a semantic model and a compliance component model.

[0041] A third aspect of the present invention provides an electronic device including a processor and a memory, wherein the processor is configured to implement the document compliance check method by executing a computer program stored in the memory.

[0042] A fourth aspect of the present invention provides a computer-readable storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, implements the aforementioned document compliance check method.

[0043] In summary, the document compliance inspection method, apparatus, electronic device, and storage medium of this invention, by performing a first preprocessing on the compliance document to be inspected, obtains an original compliance document sample set corresponding to the compliance type of the compliance document to be inspected, avoiding the acquisition of other compliance documents unrelated to the compliance type, thus ensuring the accuracy of the original compliance document sample set. Subsequently, the original compliance document sample set is considered during semantic model training, improving the accuracy of the trained semantic model, and consequently improving the accuracy of the obtained standardized compliance rules. A compliance inspection model is created by inputting the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected. The compliance inspection model includes a semantic model and a compliance component model. By recombinizing the inspection result information between the standard compliance documents and the standardized compliance rules output by the semantic model and training it in the compliance component model, the trained compliance component model becomes more accurate. Using the trained compliance component model to perform compliance inspection on the compliance document to be inspected improves the accuracy of the compliance inspection results. Attached Figure Description

[0044] Figure 1 This is a flowchart of the document compliance check method provided in Embodiment 1 of the present invention.

[0045] Figure 2 This is a structural diagram of the document compliance inspection device provided in Embodiment 2 of the present invention.

[0046] Figure 3 This is a schematic diagram of the structure of the electronic device provided in Embodiment 3 of the present invention. Detailed Implementation

[0047] To better understand the above-mentioned objects, features, and advantages of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments of the present invention and the features thereof can be combined with each other.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0049] Example 1

[0050] Figure 1 This is a flowchart of the document compliance check method provided in Embodiment 1 of the present invention.

[0051] In this embodiment, the document compliance check method can be applied to electronic devices. For electronic devices that need to perform document compliance checks, the document compliance check function provided by the method of this invention can be directly integrated into the electronic device, or it can run in the electronic device in the form of a software development kit (SDK).

[0052] like Figure 1 As shown, the document compliance check method specifically includes the following steps. Depending on different needs, the order of the steps in this flowchart can be changed, and some can be omitted.

[0053] S11, parse the received file compliance check request and obtain the compliance file to be checked.

[0054] In this embodiment, many application scenarios require compliance checks on documents. For example, in the financial sector, many traditional financial industries are undergoing retail and digital transformation, evolving from offline to online, intelligent, and self-service. This requires customized application services based on different financial business scenarios. During the customization of application services, standardized compliance rules are set up to check the compliance of documents and ensure service quality.

[0055] In this embodiment, when a user performs a document compliance check, they initiate a document compliance check request to the server through a client. Specifically, the client can be a smartphone, iPad, or other existing smart device, and the server can be a document compliance check subsystem. During the document compliance check process, the client can send a document compliance check request to the document compliance check subsystem, which is used to receive the document compliance check request sent by the client.

[0056] In an optional embodiment, parsing the received document compliance check request to obtain the compliance document to be checked includes:

[0057] Parse the document compliance check request to obtain the business type of the compliance document to be checked;

[0058] The API to be invoked is determined based on the business type.

[0059] The API call is initiated to retrieve the compliance documents to be checked from the corresponding preset database to be checked.

[0060] Specifically, the compliance documents to be inspected include text files, video files, or audio files.

[0061] For example, in the financial sector, the text file can be other documents such as the "Administrative Measures for Due Diligence of Tax-Related Information on Non-Resident Financial Accounts," the "Interim Measures for the Management of the Basic Database of Personal Credit Information," and the "Administrative Measures for RMB Bank Settlement Accounts." In this embodiment, different business types correspond to different calling interfaces. When obtaining compliance documents to be checked, the corresponding business type's calling interface is activated to retrieve the documents from the corresponding preset database to be checked, thereby improving the efficiency of obtaining compliance documents to be checked.

[0062] S12, perform a first preprocessing on the compliance document to be inspected to obtain the compliance type corresponding to the compliance document to be inspected.

[0063] In this embodiment, the first preprocessing refers to determining the corresponding compliance type based on the tag set in the compliance document to be inspected. Specifically, the compliance document to be inspected contains a tag set.

[0064] For example, if the compliance document to be inspected is NY / T 435-2021 Green Food Fruit and Vegetable Crisps, the label set can be: product, storage, classification, packaging, label, inspection rules, fruit, green food, crisps, vegetables, transportation, edible fungi.

[0065] In an optional embodiment, the first preprocessing of the compliance document to be inspected to obtain the compliance type corresponding to the compliance document to be inspected includes:

[0066] Identify the tags on the compliance documents to be inspected;

[0067] The identified labels are input into a pre-trained compliance type classification model to obtain the compliance type corresponding to the compliance document to be inspected.

[0068] In this embodiment, a classification model can be pre-trained. Specifically, the classification model training process includes: acquiring historical compliance documents and their corresponding compliance types; extracting the label set corresponding to each historical compliance document and using the label set as a sample dataset; dividing the sample dataset into a training set and a test set; inputting the label set and corresponding compliance type in the training set into a preset neural network for training to obtain a compliance type classification model; inputting the test set into the compliance type classification model for testing and calculating the test pass rate; if the test pass rate is greater than or equal to a preset pass rate threshold, determining that the compliance type classification model training has ended; if the test pass rate is less than the preset pass rate threshold, increasing the number of training sets and retraining the compliance type classification model.

[0069] In this embodiment, a compliance type classification model is pre-trained. This model represents the compliance type of the compliance document to be inspected. The compliance type can be limited to, but is not limited to, seafood, daily necessities, fruits, restaurants, and clothing. Historical compliance documents for each compliance type are extracted, and a tag set for these historical compliance documents is extracted. Then, the compliance type classification model is trained based on the extracted tag set and the compliance type. Subsequently, only the tag set of each compliance document needs to be obtained to identify the compliance type to which the compliance document belongs through the compliance type classification model, resulting in high accuracy in identifying the compliance type of the compliance document to be inspected.

[0070] S13. Based on the compliance type corresponding to the compliance document to be inspected, obtain the original compliance document sample set.

[0071] In an optional embodiment, obtaining the original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected includes:

[0072] Obtain the target data source corresponding to the compliance type of the compliance document to be inspected from the preset source database;

[0073] The web crawler technology is used to obtain a sample set of original compliant documents from the target data source according to preset rules.

[0074] In this embodiment, the target data source can be one or more, and the crawling technology is existing technology, which will not be described in detail here.

[0075] In this embodiment, since compliance rules are constantly updated and changed over time, acquisition rules can be preset. Specifically, the acquisition rules can be determined based on the time of the received file compliance check request, and the rules for acquiring the original compliance file sample set are different for different time periods.

[0076] In this embodiment, the original compliance document sample set is obtained based on the compliance type corresponding to the compliance document to be inspected, avoiding the acquisition of other compliance documents unrelated to the compliance type, thus ensuring the accuracy of the original compliance document sample set. The original compliance document sample set is then considered when training the semantic model, which improves the accuracy of the trained semantic model and, consequently, the accuracy of the obtained standardized compliance rules.

[0077] S14, perform a second preprocessing on the original compliance document sample set to obtain a standard compliance document sample set.

[0078] In this embodiment, the second preprocessing refers to standardizing the documents in the original compliance document sample set and converting them into standard compliance documents that can meet the requirements of the compliance rules.

[0079] In an optional embodiment, the second preprocessing of the original compliance document sample set to obtain the standard compliance document sample set includes:

[0080] Based on the compliance type of the compliance document to be inspected, obtain the corresponding standard-defined data structure;

[0081] According to the data structure, each original compliance document sample in the original compliance document sample set is transformed to obtain a standard compliance document sample set.

[0082] In this embodiment, the data structures of compliance documents for different compliance types are different. For the compliance type corresponding to the compliance document to be inspected, the corresponding standard-defined data structure is obtained and each original compliance document is converted to obtain a standard compliance document sample set.

[0083] S15, Create a compliance inspection model by inputting the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected. The compliance inspection model includes a semantic model and a compliance component model.

[0084] In this embodiment, the semantic model is connected in series with the compliance component model. Specifically, the input of the compliance component model depends on the output of the semantic model.

[0085] In an optional embodiment, the step of inputting the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected includes:

[0086] The standard compliance document sample set is input into the semantic model to obtain standardized compliance rules;

[0087] The standardized compliance rules and the compliance documents to be checked are input into the compliance component model connected to the semantic model to obtain the compliance check results of the compliance documents to be checked.

[0088] In this embodiment, since the semantic model is trained using a standard compliance document sample set corresponding to the compliance type, the accuracy of the trained semantic model is ensured, thereby improving the accuracy of the standardized compliance rules.

[0089] In this embodiment, each standard compliance file in the standard compliance file sample set is marked with standardized compliance rules.

[0090] Specifically, the training process of the semantic model includes:

[0091] Obtain multiple historical standard compliance documents corresponding to the compliance type and a standardized compliance rule set for each historical standard compliance document;

[0092] Find the semantic vector corresponding to each standardized compliance rule in the standardized compliance rules of each historical standard compliance document from the preset semantic table, and construct the semantic matrix of each historical standard compliance document;

[0093] The semantic matrices of the aforementioned historical standard compliance documents are determined as a semantic sample set;

[0094] The semantic sample set is divided into a training set and a test set according to a preset ratio;

[0095] The training set is input into a preset neural network for training to obtain a semantic model;

[0096] The test set is input into the semantic model for testing, and the test pass rate is calculated.

[0097] If the test pass rate is greater than or equal to a preset pass rate threshold, the semantic model training is considered complete; if the test pass rate is less than the preset pass rate threshold, the number of training sets is increased, and the semantic model is retrained.

[0098] Furthermore, the step of inputting the standardized compliance rules and the compliance document to be inspected into the compliance component model connected to the semantic model to obtain the compliance inspection result of the compliance document to be inspected includes:

[0099] The compliance document to be inspected is converted according to the data structure defined in the standard corresponding to the compliance type to obtain the standard compliance document to be inspected;

[0100] The standard compliance document to be checked and the standardized compliance rules are input into the compliance component model connected to the semantic model to obtain the compliance check result of the compliance document to be checked.

[0101] Specifically, a compliance component model can be pre-trained, wherein the training process of the compliance component model is the same as that of the semantic model, and the compliance component model is trained based on the inspection result information between each historical standard compliance rule and the corresponding historical standard compliance document.

[0102] In this embodiment, the compliance check result includes both the compliance document to be checked not meeting the compliance requirements and the compliance document to be checked meeting the compliance requirements. When the compliance document to be checked does not meet the compliance requirements, the compliance check result contains specific information about the non-compliance. For example, Article N of the "Administrative Measures for RMB Bank Settlement Accounts" states that the bank settlement date does not meet the settlement date corresponding to the standardized compliance rules.

[0103] In this embodiment, since the standardized compliance rules are identified through a semantic model, the problem of low accuracy of compliance check results is solved in the prior art, which relies on the user's experience and ability and may not be able to objectively understand the compliance rules. At the same time, the check result information between the standard compliance documents and the standardized compliance rules is recombined and trained in the compliance component model, making the trained compliance component model more accurate. The compliance check is performed on the compliance documents to be checked using the trained compliance component model, thereby improving the accuracy of the compliance check results.

[0104] For example, in the food sector, referring to the basic process of national food safety standard testing, the latest national food safety standard documents are input into the semantic model to obtain relevant standard / requirement details, i.e., standardized compliance rules; the compliance documents to be checked under the standardized compliance rules are input into the compliance component model to check whether the documents to be checked comply with the standard documents.

[0105] In summary, the document compliance check method described in this embodiment, by performing a first preprocessing on the compliance document to be checked, obtains an original compliance document sample set corresponding to the compliance type of the compliance document to be checked, avoiding the acquisition of other compliance documents unrelated to the compliance type, thus ensuring the accuracy of the original compliance document sample set. Subsequently, the original compliance document sample set is considered during semantic model training, improving the accuracy of the trained semantic model, and consequently improving the accuracy of the obtained standardized compliance rules. A compliance check model is created by inputting the standard compliance document sample set and the compliance document to be checked into the compliance check model to obtain the compliance check result of the compliance document to be checked. The compliance check model includes a semantic model and a compliance component model. By recombinizing the check result information between the standard compliance documents and the standardized compliance rules output by the semantic model and training it in the compliance component model, the trained compliance component model becomes more accurate. Using the trained compliance component model to perform compliance checks on the compliance document to be checked improves the accuracy of the compliance check results.

[0106] Example 2

[0107] Figure 2 This is a structural diagram of the document compliance inspection device provided in Embodiment 2 of the present invention.

[0108] In some embodiments, the document compliance checking device 20 may include multiple functional modules composed of program code segments. The program code of each program segment in the document compliance checking device 20 may be stored in the memory of an electronic device and executed by the at least one processor to perform (see details). Figure 1 (Description) Functionality of document compliance checks.

[0109] In this embodiment, the document compliance checking device 20 can be divided into multiple functional modules according to its functions. These functional modules may include: a parsing and acquisition module 201, a first preprocessing module 202, an acquisition module 203, a second preprocessing module 204, and a creation and input module 205. The module referred to in this invention is a series of computer-readable instruction segments that can be executed by at least one processor and perform a fixed function, stored in memory. In this embodiment, the functions of each module will be detailed in subsequent embodiments.

[0110] The parsing and acquisition module 201 is used to parse the received file compliance check request and acquire the compliance file to be checked.

[0111] In this embodiment, many application scenarios require compliance checks on documents. For example, in the financial sector, many traditional financial industries are undergoing retail and digital transformation, evolving from offline to online, intelligent, and self-service. This requires customized application services based on different financial business scenarios. During the customization of application services, standardized compliance rules are set up to check the compliance of documents and ensure service quality.

[0112] In this embodiment, when a user performs a document compliance check, they initiate a document compliance check request to the server through a client. Specifically, the client can be a smartphone, iPad, or other existing smart device, and the server can be a document compliance check subsystem. During the document compliance check process, the client can send a document compliance check request to the document compliance check subsystem, which is used to receive the document compliance check request sent by the client.

[0113] In an optional embodiment, the parsing and acquisition module 201 parses the received document compliance check request and acquires the compliance document to be checked, including:

[0114] Parse the document compliance check request to obtain the business type of the compliance document to be checked;

[0115] The API to be invoked is determined based on the business type.

[0116] The API call is initiated to retrieve the compliance documents to be checked from the corresponding preset database to be checked.

[0117] Specifically, the compliance documents to be inspected include text files, video files, or audio files.

[0118] For example, in the financial sector, the text file can be other documents such as the "Administrative Measures for Due Diligence of Tax-Related Information on Non-Resident Financial Accounts," the "Interim Measures for the Management of the Basic Database of Personal Credit Information," and the "Administrative Measures for RMB Bank Settlement Accounts." In this embodiment, different business types correspond to different calling interfaces. When obtaining compliance documents to be checked, the corresponding business type's calling interface is activated to retrieve the documents from the corresponding preset database to be checked, thereby improving the efficiency of obtaining compliance documents to be checked.

[0119] The first preprocessing module 202 is used to perform a first preprocessing on the compliance document to be inspected and obtain the compliance type corresponding to the compliance document to be inspected.

[0120] In this embodiment, the first preprocessing refers to determining the corresponding compliance type based on the tag set in the compliance document to be inspected. Specifically, the compliance document to be inspected contains a tag set.

[0121] For example, if the compliance document to be inspected is NY / T 435-2021 Green Food Fruit and Vegetable Crisps, the label set can be: product, storage, classification, packaging, label, inspection rules, fruit, green food, crisps, vegetables, transportation, edible fungi.

[0122] In an optional embodiment, the first preprocessing module 202 performs a first preprocessing on the compliance document to be inspected, and obtains the compliance type corresponding to the compliance document to be inspected, including:

[0123] Identify the tags on the compliance documents to be inspected;

[0124] The identified labels are input into a pre-trained compliance type classification model to obtain the compliance type corresponding to the compliance document to be inspected.

[0125] In this embodiment, a classification model can be pre-trained. Specifically, the classification model training process includes: acquiring historical compliance documents and their corresponding compliance types; extracting the label set corresponding to each historical compliance document and using the label set as a sample dataset; dividing the sample dataset into a training set and a test set; inputting the label set and corresponding compliance type in the training set into a preset neural network for training to obtain a compliance type classification model; inputting the test set into the compliance type classification model for testing and calculating the test pass rate; if the test pass rate is greater than or equal to a preset pass rate threshold, determining that the compliance type classification model training has ended; if the test pass rate is less than the preset pass rate threshold, increasing the number of training sets and retraining the compliance type classification model.

[0126] In this embodiment, a compliance type classification model is pre-trained. This model represents the compliance type of the compliance document to be inspected. The compliance type can be limited to, but is not limited to, seafood, daily necessities, fruits, restaurants, and clothing. Historical compliance documents for each compliance type are extracted, and a tag set for these historical compliance documents is extracted. Then, the compliance type classification model is trained based on the extracted tag set and the compliance type. Subsequently, only the tag set of each compliance document needs to be obtained to identify the compliance type to which the compliance document belongs through the compliance type classification model, resulting in high accuracy in identifying the compliance type of the compliance document to be inspected.

[0127] The acquisition module 203 is used to acquire an original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected.

[0128] In an optional embodiment, the acquisition module 203 acquires the original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected, including:

[0129] Obtain the target data source corresponding to the compliance type of the compliance document to be inspected from the preset source database;

[0130] The web crawler technology is used to obtain a sample set of original compliant documents from the target data source according to preset rules.

[0131] In this embodiment, the target data source can be one or more, and the crawling technology is existing technology, which will not be described in detail here.

[0132] In this embodiment, since compliance rules are constantly updated and changed over time, acquisition rules can be preset. Specifically, the acquisition rules can be determined based on the time of the received file compliance check request, and the rules for acquiring the original compliance file sample set are different for different time periods.

[0133] In this embodiment, the original compliance document sample set is obtained based on the compliance type corresponding to the compliance document to be inspected, avoiding the acquisition of other compliance documents unrelated to the compliance type, thus ensuring the accuracy of the original compliance document sample set. The original compliance document sample set is then considered when training the semantic model, which improves the accuracy of the trained semantic model and, consequently, the accuracy of the obtained standardized compliance rules.

[0134] The second preprocessing module 205 is used to perform a second preprocessing on the original compliance document sample set to obtain a standard compliance document sample set.

[0135] In this embodiment, the second preprocessing refers to standardizing the documents in the original compliance document sample set and converting them into standard compliance documents that can meet the requirements of the compliance rules.

[0136] In an optional embodiment, the second preprocessing module 205 performs a second preprocessing on the original compliance document sample set to obtain a standard compliance document sample set including:

[0137] Based on the compliance type of the compliance document to be inspected, obtain the corresponding standard-defined data structure;

[0138] According to the data structure, each original compliance document sample in the original compliance document sample set is transformed to obtain a standard compliance document sample set.

[0139] In this embodiment, the data structures of compliance documents for different compliance types are different. For the compliance type corresponding to the compliance document to be inspected, the corresponding standard-defined data structure is obtained and each original compliance document is converted to obtain a standard compliance document sample set.

[0140] The creation and input module 205 is used to create a compliance inspection model, input the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model, and obtain the compliance inspection result of the compliance document to be inspected. The compliance inspection model includes a semantic model and a compliance component model.

[0141] In this embodiment, the semantic model is connected in series with the compliance component model. Specifically, the input of the compliance component model depends on the output of the semantic model.

[0142] In an optional embodiment, the creation and input module 205 inputs the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected, including:

[0143] The standard compliance document sample set is input into the semantic model to obtain standardized compliance rules;

[0144] The standardized compliance rules and the compliance documents to be checked are input into the compliance component model connected to the semantic model to obtain the compliance check results of the compliance documents to be checked.

[0145] In this embodiment, since the semantic model is trained using a standard compliance document sample set corresponding to the compliance type, the accuracy of the trained semantic model is ensured, thereby improving the accuracy of the standardized compliance rules.

[0146] In this embodiment, each standard compliance file in the standard compliance file sample set is marked with standardized compliance rules.

[0147] Specifically, the training process of the semantic model includes:

[0148] Obtain multiple historical standard compliance documents corresponding to the compliance type and a standardized compliance rule set for each historical standard compliance document;

[0149] Find the semantic vector corresponding to each standardized compliance rule in the standardized compliance rules of each historical standard compliance document from the preset semantic table, and construct the semantic matrix of each historical standard compliance document;

[0150] The semantic matrices of the aforementioned historical standard compliance documents are determined as a semantic sample set;

[0151] The semantic sample set is divided into a training set and a test set according to a preset ratio;

[0152] The training set is input into a preset neural network for training to obtain a semantic model;

[0153] The test set is input into the semantic model for testing, and the test pass rate is calculated.

[0154] If the test pass rate is greater than or equal to a preset pass rate threshold, the semantic model training is considered complete; if the test pass rate is less than the preset pass rate threshold, the number of training sets is increased, and the semantic model is retrained.

[0155] Furthermore, the step of inputting the standardized compliance rules and the compliance document to be inspected into the compliance component model connected to the semantic model to obtain the compliance inspection result of the compliance document to be inspected includes:

[0156] The compliance document to be inspected is converted according to the data structure defined in the standard corresponding to the compliance type to obtain the standard compliance document to be inspected;

[0157] The standard compliance document to be checked and the standardized compliance rules are input into the compliance component model connected to the semantic model to obtain the compliance check result of the compliance document to be checked.

[0158] Specifically, a compliance component model can be pre-trained, wherein the training process of the compliance component model is the same as that of the semantic model, and the compliance component model is trained based on the inspection result information between each historical standard compliance rule and the corresponding historical standard compliance document.

[0159] In this embodiment, the compliance check result includes both the compliance document to be checked not meeting the compliance requirements and the compliance document to be checked meeting the compliance requirements. When the compliance document to be checked does not meet the compliance requirements, the compliance check result contains specific information about the non-compliance. For example, Article N of the "Administrative Measures for RMB Bank Settlement Accounts" states that the bank settlement date does not meet the settlement date corresponding to the standardized compliance rules.

[0160] In this embodiment, since the standardized compliance rules are identified through a semantic model, the problem of low accuracy of compliance check results is solved in the prior art, which relies on the user's experience and ability and may not be able to objectively understand the compliance rules. At the same time, the check result information between the standard compliance documents and the standardized compliance rules is recombined and trained in the compliance component model, making the trained compliance component model more accurate. The compliance check is performed on the compliance documents to be checked using the trained compliance component model, thereby improving the accuracy of the compliance check results.

[0161] For example, in the food sector, referring to the basic process of national food safety standard testing, the latest national food safety standard documents are input into the semantic model to obtain relevant standard / requirement details, i.e., standardized compliance rules; the compliance documents to be checked under the standardized compliance rules are input into the compliance component model to check whether the documents to be checked comply with the standard documents.

[0162] In summary, the document compliance checking device described in this embodiment, by performing a first preprocessing on the compliance document to be checked, obtains an original compliance document sample set corresponding to the compliance type of the compliance document to be checked, avoiding the acquisition of other compliance documents unrelated to the compliance type, thus ensuring the accuracy of the original compliance document sample set. Subsequently, the original compliance document sample set is considered during semantic model training, improving the accuracy of the trained semantic model, and consequently improving the accuracy of the obtained standardized compliance rules. A compliance checking model is created by inputting the standard compliance document sample set and the compliance document to be checked into the compliance checking model to obtain the compliance checking result of the compliance document to be checked. The compliance checking model includes a semantic model and a compliance component model. By recombinizing the check result information between the standard compliance documents and the standardized compliance rules output by the semantic model and training it in the compliance component model, the trained compliance component model becomes more accurate. Using the trained compliance component model to perform compliance checks on the compliance document to be checked improves the accuracy of the compliance check results.

[0163] Example 3

[0164] See Figure 3 The diagram shown is a structural schematic of an electronic device provided in Embodiment 3 of the present invention. In a preferred embodiment of the present invention, the electronic device 3 includes a memory 31, at least one processor 32, at least one communication bus 33, and a transceiver 34.

[0165] Those skilled in the art should understand that Figure 3The structure of the electronic device shown does not constitute a limitation of the embodiments of the present invention. It can be a bus structure or a star structure. The electronic device 3 may also include more or fewer other hardware or software than shown, or different component arrangements.

[0166] In some embodiments, the electronic device 3 is an electronic device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions. Its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), digital processors, and embedded devices. The electronic device 3 may also include client devices, including, but not limited to, any electronic product capable of human-computer interaction with a client via a keyboard, mouse, remote control, touchpad, or voice control device, such as personal computers, tablet computers, smartphones, and digital cameras.

[0167] It should be noted that the electronic device 3 is merely an example. Other existing or future electronic products that are suitable for this invention should also be included within the scope of protection of this invention and are incorporated herein by reference.

[0168] In some embodiments, the memory 31 is used to store program code and various data, such as the document compliance checking device 20 installed in the electronic device 3, and to achieve high-speed and automatic access to programs or data during the operation of the electronic device 3. The memory 31 includes read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0169] In some embodiments, the at least one processor 32 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits packaged with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The at least one processor 32 is the control unit of the electronic device 3, connecting various components of the entire electronic device 3 via various interfaces and lines. It executes programs or modules stored in the memory 31 and calls data stored in the memory 31 to perform various functions and process data of the electronic device 3.

[0170] In some embodiments, the at least one communication bus 33 is configured to enable communication between the memory 31 and the at least one processor 32, etc.

[0171] Although not shown, the electronic device 3 may also include a power supply (such as a battery) to power the various components. Optionally, the power supply may be logically connected to the at least one processor 32 via a power management device, thereby enabling functions such as charging, discharging, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 3 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0172] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0173] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) or processor to execute portions of the methods described in the various embodiments of the present invention.

[0174] In a further embodiment, combined with Figure 2 The at least one processor 32 can execute the operating device of the electronic device 3 and various installed applications (such as the document compliance checking device 20), program code, etc., for example, the various modules mentioned above.

[0175] The memory 31 stores program code, and the at least one processor 32 can call the program code stored in the memory 31 to execute related functions. For example, Figure 2 The modules described herein are program codes stored in the memory 31 and executed by the at least one processor 32, thereby realizing the functions of the modules to achieve the purpose of document compliance checking.

[0176] For example, the program code can be divided into one or more modules / units, which are stored in the memory 31 and executed by the processor 32 to complete this application. The one or more modules / units can be a series of computer-readable instruction segments capable of performing a specific function, which describe the execution process of the program code in the electronic device 3. For example, the program code can be divided into a parsing and acquisition module 201, a first preprocessing module 202, an acquisition module 203, a second preprocessing module 204, and a creation and input module 205.

[0177] In one embodiment of the present invention, the memory 31 stores a plurality of computer-readable instructions, which are executed by the at least one processor 32 to perform a document compliance check function.

[0178] Specifically, the specific implementation method of the above instructions by the at least one processor 32 can be referred to Figure 1 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0179] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0180] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0181] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0182] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other elements, and the singular does not exclude the plural. Multiple elements or devices recited in the present invention may also be implemented by a single element or device in software or hardware. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0183] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for checking document compliance, characterized in that, The method includes: Parse the received document compliance check request to obtain the compliance document to be checked; The compliance document to be inspected undergoes a first preprocessing step to obtain the compliance type corresponding to the compliance document to be inspected, including: identifying the label of the compliance document to be inspected; and inputting the identified label into a pre-trained compliance type classification model to obtain the compliance type corresponding to the compliance document to be inspected. Based on the compliance type corresponding to the compliance documents to be inspected, obtain the original compliance document sample set; The original compliance document sample set is subjected to a second preprocessing step to obtain a standard compliance document sample set; A compliance inspection model comprising a semantic model and a compliance component model is created. The standard compliance document sample set and the compliance document to be inspected are input into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected. This includes: inputting the standard compliance document sample set into the semantic model to obtain standardized compliance rules; and inputting the standardized compliance rules and the compliance document to be inspected into a compliance component model connected to the semantic model to obtain the compliance inspection result of the compliance document to be inspected. The semantic model and the compliance component model are concatenated, and the input of the compliance component model depends on the output of the semantic model.

2. The document compliance check method as described in claim 1, characterized in that, The process of obtaining the original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected includes: Obtain the target data source corresponding to the compliance type of the compliance document to be inspected from the preset source database; The web crawler technology is used to obtain a sample set of original compliant documents from the target data source according to preset rules.

3. The document compliance check method as described in claim 1, characterized in that, The second preprocessing of the original compliance document sample set to obtain the standard compliance document sample set includes: Based on the compliance type of the compliance document to be inspected, obtain the corresponding standard-defined data structure; According to the data structure, each original compliance document sample in the original compliance document sample set is transformed to obtain a standard compliance document sample set.

4. The document compliance check method as described in claim 1, characterized in that, The training process of the semantic model includes: Obtain multiple historical standard compliance documents corresponding to the compliance type and a standardized compliance rule set for each historical standard compliance document; Find the semantic vector corresponding to each standardized compliance rule in the standardized compliance rules of each historical standard compliance document from the preset semantic table, and construct the semantic matrix of each historical standard compliance document; The semantic matrices of the aforementioned historical standard compliance documents are determined as a semantic sample set; The semantic sample set is divided into a training set and a test set according to a preset ratio; The training set is input into a preset neural network for training to obtain a semantic model; The test set is input into the semantic model for testing, and the test pass rate is calculated. If the test pass rate is greater than or equal to a preset pass rate threshold, the semantic model training is considered complete; if the test pass rate is less than the preset pass rate threshold, the number of training sets is increased, and the semantic model is retrained.

5. The document compliance check method as described in claim 1, characterized in that, The step of inputting the standardized compliance rules and the compliance document to be inspected into the compliance component model connected to the semantic model to obtain the compliance inspection result of the compliance document to be inspected includes: The compliance document to be inspected is converted according to the data structure defined in the standard corresponding to the compliance type to obtain the standard compliance document to be inspected; The standard compliance document to be checked and the standardized compliance rules are input into the compliance component model connected to the semantic model to obtain the compliance check result of the compliance document to be checked.

6. A document compliance checking device, characterized in that, The device includes: The parsing and retrieval module is used to parse the received file compliance check requests and retrieve the compliance files to be checked; The first preprocessing module is used to perform a first preprocessing on the compliance document to be inspected to obtain the compliance type corresponding to the compliance document to be inspected, including: identifying the label of the compliance document to be inspected; inputting the identified label into a pre-trained compliance type classification model to obtain the compliance type corresponding to the compliance document to be inspected; The acquisition module is used to acquire an original compliance document sample set based on the compliance type corresponding to the compliance document to be inspected; The second preprocessing module is used to perform a second preprocessing on the original compliance document sample set to obtain a standard compliance document sample set. A creation and input module is used to create a compliance inspection model that includes a semantic model and a compliance component model. The module inputs the standard compliance document sample set and the compliance document to be inspected into the compliance inspection model to obtain the compliance inspection result of the compliance document to be inspected. This includes: inputting the standard compliance document sample set into the semantic model to obtain standardized compliance rules; and inputting the standardized compliance rules and the compliance document to be inspected into a compliance component model connected to the semantic model to obtain the compliance inspection result of the compliance document to be inspected. The semantic model and the compliance component model are connected in series, and the input of the compliance component model depends on the output of the semantic model.

7. An electronic device, characterized in that, The electronic device includes a processor and a memory, the processor being configured to implement the document compliance check method as described in any one of claims 1 to 5 when executing a computer program stored in the memory.

8. A computer-readable storage medium storing a computer program thereon, characterized in that, When the computer program is executed by the processor, it implements the document compliance check method as described in any one of claims 1 to 5.

Images