CN115471307A - Audit evaluation information generation method and device based on knowledge graph and electronic equipment - Google Patents

Abstract

The embodiment of the disclosure discloses a knowledge graph-based audit assessment information generation method and device and electronic equipment. One embodiment of the method comprises: acquiring a target audit report text; performing text cleaning on the target audit report text to generate text information; generating a target knowledge graph corresponding to the text information; determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set; generating audit evaluation information corresponding to a target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model; and sending the audit evaluation information to a target display terminal for display. The embodiment improves the auditing efficiency and the accuracy of the generated auditing assessment information.

Description

Audit evaluation information generation method and device based on knowledge graph and electronic equipment

Technical Field

The embodiment of the disclosure relates to the technical field of computers, in particular to a knowledge graph-based audit evaluation information generation method and device and electronic equipment.

Background

The auditing means a technical means for examining an auditing report according to auditing rules. At present, when auditing reports, the following methods are generally adopted: and auditing the audit report in a manual mode to generate corresponding audit evaluation information.

However, when the above-described manner is adopted, there are often technical problems as follows:

firstly, with the increase of the number of the audit reports, the audit reports are difficult to audit in time in a manual mode, so that the audit efficiency is low;

and secondly, auditing is performed in a manual mode, and the accuracy of generated audit evaluation information cannot be guaranteed due to the fact that the audit is performed by people's experience.

Disclosure of Invention

This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Some embodiments of the present disclosure propose a method, an apparatus, and an electronic device for generating audit trail information based on a knowledge graph to solve one or more of the technical problems mentioned in the above background.

In a first aspect, some embodiments of the present disclosure provide a method for generating audit trail information based on a knowledge-graph, the method including: acquiring a target audit report text; performing text cleaning on the target audit report text to generate text information; generating a target knowledge graph corresponding to the text information; determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set; generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model; and sending the audit evaluation information to a target display terminal for display.

Optionally, the text cleaning of the target audit report text to generate text information includes: reading the target audit report text to generate candidate text information; performing entity identification on the candidate text information to generate a first entity information set, wherein first entity information in the first entity information set includes: entity location information and entity information; determining entity similarity between the entity information included in the first entity information and each entity information in the target entity library, to generate a similarity value; screening out entity information of which the corresponding similarity value is positioned in a target interval from the target entity library, and taking the entity information as candidate entity information to obtain a candidate entity information set; performing entity replacement on the first entity information included in the candidate text information according to entity position information and target candidate entity information included in the first entity information, wherein the target candidate entity information is candidate entity information meeting a first screening condition in the candidate entity information set; and determining the candidate text information after the entity replacement as the text information.

Optionally, generating a target knowledge graph corresponding to the text information includes: carrying out entity identification on the text information to generate a second entity information set; entity removing is carried out on second entity information in the second entity information set to generate removed entity information, and a removed entity information set is obtained; determining the corresponding relation information of each entity information in the entity information set after the elimination to obtain a relation information group set; and generating the target knowledge graph according to the removed entity information set and the relation information group set.

Optionally, determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set, including: screening a knowledge graph meeting second screening conditions from a knowledge graph library to serve as a candidate knowledge graph to obtain a candidate knowledge graph set, wherein the second screening conditions are that the number of entity nodes contained in the knowledge graph is greater than or equal to the number of entity nodes contained in the target knowledge graph, and the number of relation edges contained in the knowledge graph is greater than or equal to the number of relation edges contained in the target knowledge graph; executing the following processing steps on each entity node in the target knowledge graph: constructing a feature vector of the entity node to generate a first entity node feature vector; constructing a feature vector for at least one relationship edge connected with the entity node to generate a first relationship edge feature vector; for each candidate knowledge-graph in the set of candidate knowledge-graphs, performing the following processing steps: constructing a feature vector for each entity node in the candidate knowledge graph to generate a second entity node feature vector to obtain a second entity node feature vector sequence; constructing a feature vector for each target relationship edge group in the candidate knowledge graph to generate a second relationship edge feature vector sequence, wherein the target relationship edge group is at least one relationship edge connected with the entity nodes of the candidate knowledge graph; determining the feature similarity of each first entity node feature vector in the obtained first entity node feature vector sequence and each second entity node feature vector in the second entity node feature vector sequence to obtain an entity node similarity value group sequence; determining the feature similarity of each first relation edge feature vector in the obtained first relation edge feature vector sequence and each second relation edge feature vector in the second relation edge feature vector sequence to obtain a relation edge similarity value group sequence; screening out the entity node similarity value with the maximum value from each entity node similarity value group in the entity node similarity value group sequence as a first candidate value to obtain a first candidate value set; screening out a relation side similarity value with the largest value from each relation side similarity value group in the relation side similarity value group sequence, and taking the relation side similarity value as a second candidate value to obtain a second candidate value group; performing weighted summation on a first candidate value in the first candidate value set and a second candidate value in the second candidate value set to generate a map similarity value corresponding to the candidate knowledge map; and screening out the candidate knowledge graphs with the corresponding graph similarity numerical values meeting the third screening condition from the candidate knowledge graph set to serve as isomorphic knowledge graphs, and obtaining the isomorphic knowledge graph set.

Optionally, the quality assessment model comprises: the method comprises the steps of collecting a feature extraction model, a first classification model, a feature fusion model and a second classification model; and generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model, wherein the audit evaluation information comprises: inputting the objective knowledge graph and the isomorphic knowledge graph in the isomorphic knowledge graph set into the feature extraction model in the feature extraction model set in parallel to generate a first feature graph feature information set and a second graph feature information set; inputting the first feature map feature information and the second feature map feature information set into the first classification model to generate a classification result; inputting the first feature map feature information and at least one second map feature information corresponding to the classification result into the feature fusion model to generate third map feature information; and inputting the third map feature information into the second classification model to generate the audit evaluation information.

Optionally, before the sending the audit evaluation information to a target display terminal for display, the method further includes: sending a communication request to a communication terminal connected with the target display terminal; in response to receiving a communication connection request sent by the communication terminal, establishing an end-to-end communication link with the communication terminal; and in response to the fact that the communication link is successfully established, transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link.

Optionally, before the transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link, the method further includes: constructing an encryption key according to the communication address and the sending time stamp of the communication terminal; and encrypting the audit evaluation information through the encryption key to generate the encrypted text.

In a second aspect, some embodiments of the present disclosure provide an apparatus for generating audit trail information based on a knowledge-graph, the apparatus comprising: an obtaining unit configured to obtain a target audit report text; the text cleaning unit is configured to perform text cleaning on the target audit report text to generate text information; the first generating unit is configured to generate a target knowledge graph corresponding to the text information; the determining unit is configured to determine an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set; a second generating unit, configured to generate audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model; and the display unit is configured to send the audit evaluation information to a target display terminal for displaying.

Optionally, the text washing unit is further configured to: reading the target audit report text to generate candidate text information; performing entity identification on the candidate text information to generate a first entity information set, wherein first entity information in the first entity information set includes: entity location information and entity information; determining entity similarity between the entity information included in the first entity information and each entity information in the target entity library to generate a similarity value; screening entity information of which the corresponding similarity value is located in a target interval from the target entity library, and taking the entity information as candidate entity information to obtain a candidate entity information set; performing entity replacement on the first entity information included in the candidate text information according to entity position information and target candidate entity information included in the first entity information, wherein the target candidate entity information is candidate entity information meeting a first screening condition in the candidate entity information set; and determining the candidate text information after the entity replacement as the text information.

Optionally, the first generating unit is further configured to: carrying out entity identification on the text information to generate a second entity information set; entity removing is carried out on second entity information in the second entity information set to generate removed entity information, and a removed entity information set is obtained; determining the relationship information corresponding to each entity information in the entity information set after the elimination to obtain a relationship information group set; and generating the target knowledge graph according to the removed entity information set and the relation information group set.

Optionally, the determining unit is further configured to: screening a knowledge graph meeting second screening conditions from a knowledge graph library to serve as a candidate knowledge graph to obtain a candidate knowledge graph set, wherein the second screening conditions are that the number of entity nodes contained in the knowledge graph is greater than or equal to the number of entity nodes contained in the target knowledge graph, and the number of relation edges contained in the knowledge graph is greater than or equal to the number of relation edges contained in the target knowledge graph; executing the following processing steps on each entity node in the target knowledge graph: constructing a feature vector of the entity node to generate a first entity node feature vector; constructing a feature vector for at least one relationship edge connected with the entity node to generate a first relationship edge feature vector; for each candidate knowledge-graph in the set of candidate knowledge-graphs, performing the following processing steps: constructing a feature vector for each entity node in the candidate knowledge graph to generate a second entity node feature vector to obtain a second entity node feature vector sequence; constructing a feature vector for each target relationship edge group in the candidate knowledge graph to generate a second relationship edge feature vector sequence, wherein the target relationship edge group is at least one relationship edge connected with the entity nodes of the candidate knowledge graph; determining the feature similarity of each first entity node feature vector in the obtained first entity node feature vector sequence and each second entity node feature vector in the second entity node feature vector sequence to obtain an entity node similarity value group sequence; determining the feature similarity of each first relation edge feature vector in the obtained first relation edge feature vector sequence and each second relation edge feature vector in the second relation edge feature vector sequence to obtain a relation edge similarity value group sequence; screening out the entity node similarity value with the maximum value from each entity node similarity value group in the entity node similarity value group sequence as a first candidate value to obtain a first candidate value set; screening out a relation side similarity value with the largest value from each relation side similarity value group in the relation side similarity value group sequence, and taking the relation side similarity value as a second candidate value to obtain a second candidate value group; carrying out weighted summation on a first candidate value in the first candidate value set and a second candidate value in the second candidate value set so as to generate a map similarity numerical value corresponding to the candidate knowledge map; and screening out the candidate knowledge graphs with the corresponding graph similarity numerical values meeting the third screening condition from the candidate knowledge graph set to serve as isomorphic knowledge graphs, and obtaining the isomorphic knowledge graph set. In practical situations, knowledge maps corresponding to different audit report texts often contain a part of different feature information. Therefore, by determining the isomorphic knowledge graph set corresponding to the target knowledge graph, the richness of the characteristics can be increased, and the generated audit evaluation information can be more accurate. Therefore, how to determine the isomorphic knowledge graph set corresponding to the target knowledge graph is particularly important. First, the isomorphic knowledge graph should contain similar number of relationship edges and number of entity nodes. Therefore, the present disclosure coarsely screens the knowledge-graph through the second screening condition to filter out useless knowledge-graphs. In addition, the graph structures of the isomorphic knowledge graph should be similar, so the method and the system determine the isomorphic knowledge graph in a mode of constructing the feature vectors of the entity nodes and the relation edges connected with the entity nodes and calculating the similarity according to the constructed feature vectors. By the method, the accuracy of the obtained isomorphic knowledge graph is greatly improved. Therefore, the accuracy of the generated audit evaluation information is improved.

Optionally, the quality assessment model comprises: the method comprises the steps of collecting a feature extraction model, a first classification model, a feature fusion model and a second classification model; and the second generating unit is further configured to: inputting the objective knowledge graph and the isomorphic knowledge graph in the isomorphic knowledge graph set into the feature extraction model in the feature extraction model set in parallel to generate a first feature graph feature information set and a second graph feature information set; inputting the first feature map feature information and the second feature map feature information set into the first classification model to generate a classification result; inputting the first feature map feature information and at least one second map feature information corresponding to the classification result into the feature fusion model to generate third map feature information; and inputting the third map feature information into the second classification model to generate the audit evaluation information.

Optionally, before the sending the audit evaluation information to a target display terminal for displaying, the apparatus further includes: sending a communication request to a communication terminal connected with the target display terminal; in response to receiving a communication connection request sent by the communication terminal, establishing an end-to-end communication link with the communication terminal; and in response to the fact that the communication link is successfully established, transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link.

Optionally, before the transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link, the apparatus further includes: constructing an encryption key according to the communication address and the sending time stamp of the communication terminal; and encrypting the audit evaluation information through the encryption key to generate the encrypted text.

In a third aspect, some embodiments of the present disclosure provide an electronic device, comprising: one or more processors; a storage device having one or more programs stored thereon, which when executed by one or more processors, cause the one or more processors to implement the method described in any of the implementations of the first aspect.

In a fourth aspect, some embodiments of the disclosure provide a computer readable medium on which a computer program is stored, wherein the program when executed by a processor implements the method described in any implementation of the first aspect.

The above embodiments of the present disclosure have the following advantages: by the aid of the method for generating the audit evaluation information based on the knowledge graph, the audit efficiency is greatly improved on the premise that accuracy of the generated audit evaluation information is guaranteed. Specifically, the reasons why the audit efficiency is low and the accuracy of the generated audit evaluation information cannot be ensured are that: firstly, with the increase of the number of the audit reports, the audit reports are difficult to audit in time in a manual mode, so that the audit efficiency is low. And secondly, auditing is performed in a manual mode, and the accuracy of generated audit evaluation information cannot be guaranteed due to the fact that the audit is performed by people's experience. Based on this, the method for generating audit assessment information based on knowledge graph according to some embodiments of the present disclosure first obtains a target audit report text. And secondly, performing text cleaning on the target audit report text to generate text information. And removing text contents irrelevant to entity extraction and relation extraction through text cleaning. And then, generating a target knowledge graph corresponding to the text information. Therefore, the entities contained in the target audit report text and the relationship among the entities are extracted. And then, determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set. And acquiring the isomorphic knowledge graph to determine the knowledge graph related to the target knowledge graph. And further generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model. By means of the isomorphic knowledge graph set, the graph characteristics of the target knowledge graph are enriched, and therefore the accuracy of the generated audit evaluation information is improved. And finally, sending the audit evaluation information to a target display terminal for display. By the method, the problems existing in manual auditing are solved, and auditing efficiency and accuracy of generated auditing assessment information are greatly improved.

Drawings

The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by referring to the following detailed description when taken in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements. It should be understood that the drawings are schematic and that elements and elements are not necessarily drawn to scale.

FIG. 1 is a schematic illustration of an application scenario of a knowledge-graph based audit trail information generation method of some embodiments of the present disclosure;

FIG. 2 is a flow diagram of some embodiments of a knowledge-graph based audit trail information generation method according to the present disclosure;

FIG. 3 is a flow diagram of further embodiments of a knowledge-graph based audit trail information generation method according to the present disclosure;

FIG. 4 is a schematic block diagram of some embodiments of a knowledge-graph based audit trail information generation apparatus according to the present disclosure;

FIG. 5 is a schematic structural diagram of an electronic device suitable for use in implementing some embodiments of the present disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is to be understood that the disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the disclosure are for illustration purposes only and are not intended to limit the scope of the disclosure.

It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.

It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence of the functions performed by the devices, modules or units.

It is noted that references to "a", "an", and "the" modifications in this disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that "one or more" may be used unless the context clearly dictates otherwise.

The names of messages or information exchanged between devices in the embodiments of the present disclosure are for illustrative purposes only, and are not intended to limit the scope of the messages or information.

The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

FIG. 1 is a schematic diagram of one application scenario of a knowledge-graph based audit trail information generation method of some embodiments of the present disclosure.

In the application scenario of fig. 1, first, computing device 101 may obtain target audit report text 102; then, the computing device 101 may perform text cleaning on the target audit report text 102 to generate text information 103; next, the computing device 101 may generate a target knowledge-graph 104 corresponding to the text information 103; further, the computing device 101 may determine an isomorphic knowledge graph corresponding to the target knowledge graph 104 to obtain an isomorphic knowledge graph set 105; in addition, the computing device 101 may generate audit evaluation information 107 corresponding to the target audit report text 102 according to the target knowledge graph 104, the isomorphic knowledge graph set 105, and the pre-trained quality evaluation model 106; finally, the computing device 101 may send the audit trail information 107 described above to the target display terminal 108 for display.

The computing device 101 may be hardware or software. When the computing device is hardware, it may be implemented as a distributed cluster composed of multiple servers or terminal devices, or may be implemented as a single server or a single terminal device. When the computing device is embodied as software, it may be installed in the hardware devices enumerated above. It may be implemented, for example, as multiple software or software modules to provide distributed services, or as a single software or software module. And is not particularly limited herein.

It should be understood that the number of computing devices in FIG. 1 is merely illustrative. There may be any number of computing devices, as implementation needs dictate.

With continued reference to FIG. 2, a flow 200 of some embodiments of a knowledge-graph based audit trail information generation method according to the present disclosure is shown. The audit evaluation information generation method based on the knowledge graph comprises the following steps:

step 201, obtaining a target audit report text.

In some embodiments, an executing entity (e.g., computing device 101 shown in fig. 1) of the knowledge-graph-based audit trail information generation method may obtain the target audit report text by means of a wired connection or a wireless connection. The target audit report text can be an audit report text to be audited and evaluated.

Step 202, text cleaning is carried out on the target audit report text to generate text information.

In some embodiments, the execution subject may perform text washing on the target audit report text to generate text information. For example, the execution subject may remove information other than the entity and the entity correspondence in the audit report text to generate the text information. For another example, the execution subject may remove punctuation marks, tab marks, and other symbols in the audit report text to generate the text information.

And step 203, generating a target knowledge graph corresponding to the text information.

In some embodiments, the execution agent may generate a target knowledge-graph corresponding to the text message. The target knowledge graph may be a graph formed according to the entity included in the text information and the relationship corresponding to the entity. The data structure of the target knowledge graph is a graph structure.

As an example, first, the execution subject may perform entity extraction and relationship extraction on the text information. Then, the executing body may construct the target knowledge graph by using the entity as a node and the relationship corresponding to the entity as an edge according to the extracted entity and entity corresponding relationship.

And 204, determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set.

In some embodiments, the executing entity may determine an isomorphic knowledge graph corresponding to the target knowledge graph, and obtain an isomorphic knowledge graph set. The isomorphic knowledge graph in the isomorphic knowledge graph set comprises nodes similar to the target knowledge graph, and the graph structure of the isomorphic knowledge graph is similar to that of the target knowledge graph.

As an example, the determining, by the execution subject, an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set may include the following steps:

the first step is to screen out the knowledge graph satisfying graph screening conditions from a knowledge graph library to serve as a preprocessing knowledge graph, and a preprocessing knowledge graph set is obtained.

The map screening condition is that the knowledge map comprises a target number of target nodes. The target nodes are the same nodes contained in the knowledge graph library and contained in the target knowledge graph. The target number may characterize a minimum number of nodes that a knowledge graph in the knowledge graph repository contains that are identical to the target knowledge graph. The target number may be determined by the following formula:

TN＝TP×TGN

where TN represents the above-mentioned target number. TP represents the target percentage. TGN represents the number of nodes in the target knowledge-graph. The target percentage is a ratio of the number of target nodes to the number of nodes in the target knowledge-graph.

Second, for each preprocessing knowledge-graph in the set of preprocessing knowledge-graphs, performing the following sub-processing steps:

the first substep, selecting a target node.

Wherein the target node is a node included in both the preprocessing knowledge-graph and the target knowledge-graph.

And a second sub-step of performing tree structure conversion on the preprocessed knowledge graph and the target knowledge graph to generate a first multi-way tree and a second multi-way tree, respectively, with the target node as a root node.

Wherein the first multi-way tree is a tree corresponding to the preprocessed knowledge graph. The second multi-way tree is a tree corresponding to the target knowledge-graph.

And a third sub-step of determining whether the first multi-way tree and the second multi-way tree are similar.

A fourth substep of determining said preprocessed knowledge-graph as a homogenous knowledge-graph in response to determining said first multi-way tree is similar to said second multi-way tree.

And step 205, generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and the pre-trained quality evaluation model.

In some embodiments, the execution subject may generate audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the set of homogeneous knowledge graphs, and the pre-trained quality evaluation model. The quality evaluation model may be a model for generating audit evaluation information corresponding to the target audit report text. The audit evaluation information can be evaluation information for representing whether the target audit report text conforms to the audit rule.

As an example, first, the executing body may perform feature extraction on the target knowledge graph to generate a first feature. Then, the executing body may perform feature extraction on each isomorphic knowledge graph in the set of isomorphic knowledge graphs to generate a second feature, resulting in a second feature set. Wherein the first feature and the second feature in the second feature set have the same feature length. Next, the execution subject may perform feature superposition on the first feature and each second feature in the second feature set to generate a superposition feature. Further, the execution body may divide each value in the superimposed feature by the target value to generate the target feature. Wherein the target value is the number of second features in the second feature set plus 1. Finally, the execution subject may input the target feature into the quality assessment model to generate the audit trail information. For example, the quality assessment model may be an LSTM (Long Short Term Memory) model.

And step 206, sending the audit evaluation information to a target display terminal for displaying.

In some embodiments, the execution subject may send the audit evaluation information to the target display terminal for display by a wired connection or a wireless connection. The target display terminal may be a terminal having a display function. For example, the target display terminal may be a computer, and the target display terminal may also be a handheld terminal having a display function.

The above embodiments of the present disclosure have the following advantages: by the aid of the method for generating the audit evaluation information based on the knowledge graph, the audit efficiency is greatly improved on the premise that accuracy of the generated audit evaluation information is guaranteed. Specifically, the reasons why the audit efficiency is low and the accuracy of the generated audit evaluation information cannot be guaranteed are that: firstly, with the increase of the number of the audit reports, the audit reports are difficult to audit in time in a manual mode, so that the audit efficiency is low. And secondly, auditing is performed in a manual mode, and the accuracy of generated audit evaluation information cannot be guaranteed due to the fact that the audit depends on the experience of people. Based on this, the method for generating audit assessment information based on knowledge graph according to some embodiments of the present disclosure first obtains a target audit report text. And secondly, performing text cleaning on the target audit report text to generate text information. And removing text contents irrelevant to entity extraction and relation extraction through text cleaning. And then, generating a target knowledge graph corresponding to the text information. Therefore, the entities contained in the target audit report text and the relationship among the entities are extracted. And then, determining the isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set. And acquiring the isomorphic knowledge graph to determine the knowledge graph related to the target knowledge graph. And further generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model. By means of the isomorphic knowledge graph set, the graph characteristics of the target knowledge graph are enriched, and therefore the accuracy of the generated audit evaluation information is improved. And finally, sending the audit evaluation information to a target display terminal for display. By the method, the problems existing in manual auditing are solved, and auditing efficiency and accuracy of generated auditing assessment information are greatly improved.

With further reference to FIG. 3, a flow 300 of further embodiments of a knowledge-graph based audit trail information generation method is illustrated. The process 300 of the audit evaluation information generation method based on the knowledge graph comprises the following steps:

step 301, obtaining a target audit report text.

In some embodiments, the specific implementation of step 301 and the technical effect thereof may refer to step 201 in the embodiment corresponding to fig. 2, and are not described herein again.

And step 302, performing text cleaning on the target audit report text to generate text information.

In some embodiments, an executing subject of the knowledge-graph-based audit assessment information generation method (e.g., computing device 101 shown in fig. 1) performs text cleansing on target audit report text to generate textual information, which may include the steps of:

and step one, reading the target audit report text to generate candidate text information.

As an example, the executing entity may read the content of the target audit report text at one time in a full reading manner to generate the candidate text information. For example, the execution subject may read the target audit report text by:

file＝open("filename")

Inf＝file.read()

wherein, the "filename" represents the file name of the target audit report text. "Inf" represents the above text information. The execution body can read the content of the target audit report text at one time through a read () method.

And secondly, performing entity recognition on the candidate text information to generate a first entity information set.

The execution subject can perform entity recognition on the candidate text information through an entity recognition model. The first entity information in the first entity information set includes: entity location information and entity information. The entity position information represents the position of the entity corresponding to the first entity information in the candidate text information. The entity information represents the entity of the corresponding position of the entity position information.

For example, the first entity information may be [ entity location information: (12,23), entity information: "Zhang three" ].

Thirdly, for each first entity information in the first entity information set, executing the following processing steps:

the first substep is to determine the entity similarity between the entity information included in the first entity information and each entity information in the target entity library to generate a similarity value.

The execution subject may generate the similarity value by determining a cosine similarity between the entity information included in the first entity information and each entity information in the target entity library.

And a second substep of screening out entity information of which the corresponding similarity value is positioned in the target interval from the target entity library, and taking the entity information as candidate entity information to obtain a candidate entity information set.

The target interval may be [100%,85% ].

And a third substep of performing entity replacement on the first entity information included in the candidate text information according to the entity location information and the target candidate entity information included in the first entity information.

The execution body may replace an entity corresponding to the first entity information at the position corresponding to the entity position information with an entity corresponding to the target candidate entity information. The target candidate entity information is candidate entity information satisfying a first filtering condition in the candidate entity information set. The first screening condition is that the similarity value corresponding to the candidate entity information is maximum.

And a fourth substep of determining the candidate text information after the entity replacement as the text information.

And step 303, generating a target knowledge graph corresponding to the text information.

In some embodiments, the generating of the target knowledge-graph corresponding to the text information by the execution subject may include the following steps:

firstly, entity recognition is carried out on the text information to generate a second entity information set.

And the second entity information in the second entity information set represents the entity in the text information. For example, the execution agent may perform entity recognition on the text message by using a (Hidden Markov Model) HMM.

And secondly, performing entity elimination on the second entity information in the second entity information set to generate eliminated entity information and obtain an eliminated entity information set.

And the entity information after being removed in the entity information set after being removed is the information corresponding to the entity related to the audit.

And thirdly, determining the corresponding relation information of each entity information in the entity information set after the elimination to obtain a relation information group set.

The relationship information groups in the relationship information group set represent a plurality of relationships corresponding to one entity information. The execution body may determine, through a BERT (Bidirectional Encoder based on transforms) model, relationship information corresponding to each entity information in the removed entity information set, to obtain a relationship information group set.

And fourthly, generating the target knowledge graph according to the entity information set after the elimination and the relation information group set.

And 304, determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set.

In some embodiments, the determining, by the execution subject, an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set may include the following steps:

and step one, screening the knowledge graph which meets the second screening condition from the knowledge graph library to serve as a candidate knowledge graph, and obtaining a candidate knowledge graph set.

The second filtering condition is that the number of entity nodes included in the knowledge graph is greater than or equal to the number of entity nodes included in the target knowledge graph, and the number of relationship edges included in the knowledge graph is greater than or equal to the number of relationship edges included in the target knowledge graph. The knowledge graph library is a library for storing knowledge graphs from multiple sources.

Secondly, executing the following processing steps on each entity node in the target knowledge graph:

the first substep, construct the eigenvector to the above-mentioned entity node, in order to produce the first entity node eigenvector.

As an example, the execution subject may perform one-hot encoding on the entity node to generate the first entity node feature vector.

As another example, the executing entity may perform feature extraction on the entity node through a convolutional neural network to generate the first entity node feature vector.

And a second substep of constructing a feature vector for at least one relationship edge connected with the entity node to generate a first relationship edge feature vector.

As an example, first, the main body may generate a plurality of feature vectors by performing one-hot encoding on at least one relationship edge connected to the entity node. Then, the execution subject may perform vector concatenation on the plurality of feature vectors to generate the first relationship edge feature vector.

Thirdly, for each candidate knowledge-graph in the candidate knowledge-graph set, executing the following processing steps:

and a first substep, constructing a feature vector for each entity node in the candidate knowledge graph to generate a second entity node feature vector, so as to obtain a second entity node feature vector sequence.

As an example, the execution subject may perform one-hot encoding on each entity node in the candidate knowledge-graph to generate a second entity node feature vector, resulting in a second entity node feature vector sequence.

And a second sub-step, constructing a feature vector for each target relation edge group in the candidate knowledge graph to generate a second relation edge feature vector sequence.

And the target relation edge group is at least one relation edge connected with the entity nodes of the candidate knowledge graph.

As an example, the executing entity may perform a one-hot encoding process on each target relationship edge group in the candidate knowledge-graph to generate a second relationship edge feature vector sequence.

As yet another example, the executing entity may perform feature extraction on each target relationship edge group in the candidate knowledge-graph through a convolutional neural network to generate a second relationship edge feature vector sequence.

And a third substep, determining the feature similarity of each first entity node feature vector in the obtained first entity node feature vector sequence and each second entity node feature vector in the second entity node feature vector sequence, and obtaining an entity node similarity value group sequence.

The feature similarity of the first entity node feature vector and the second entity node feature vector may be cosine similarity.

And a fourth substep of determining the feature similarity between each first relationship edge feature vector in the obtained first relationship edge feature vector sequence and each second relationship edge feature vector in the second relationship edge feature vector sequence to obtain a relationship edge similarity value group sequence.

The feature similarity of the first relationship edge feature vector and the second relationship edge feature vector may be a cosine similarity.

And a fifth substep of screening out the entity node similarity value with the largest value from each entity node similarity value group in the entity node similarity value group sequence as a first candidate value to obtain a first candidate value set.

And a sixth substep, selecting the relationship edge similarity value with the largest value from each relationship edge similarity value group in the relationship edge similarity value group sequence as a second candidate value, and obtaining a second candidate value group.

And a seventh sub-step of performing weighted summation on a first candidate value in the first candidate value set and a second candidate value in the second candidate value set to generate a map similarity value corresponding to the candidate knowledge map.

And fourthly, screening out the corresponding candidate knowledge graph with the graph similarity value meeting the third screening condition from the candidate knowledge graph set to serve as the isomorphic knowledge graph, and obtaining the isomorphic knowledge graph set.

Wherein the third screening condition is that the map similarity value is equal to or greater than the map similarity value of the target position in the obtained map similarity value sequence. The numerical sequence of the map similarity is ordered from large to small. The target position may be "4".

In practical situations, knowledge maps corresponding to different audit report texts often contain a part of different feature information. Therefore, by determining the isomorphic knowledge graph set corresponding to the target knowledge graph, the richness of the characteristics can be increased, and the generated audit evaluation information can be more accurate. Therefore, how to determine the isomorphic knowledge graph set corresponding to the target knowledge graph is particularly important. First, the isomorphic knowledge graph should contain similar number of relationship edges and number of entity nodes. Therefore, the present disclosure performs a coarse screening of the knowledge-graph by the second screening condition to filter out useless knowledge-graphs. In addition, the graph structures of the isomorphic knowledge graph should be similar, so the method and the system determine the isomorphic knowledge graph in a mode of constructing the feature vectors of the entity nodes and the relation edges connected with the entity nodes and calculating the similarity according to the constructed feature vectors. By the method, the accuracy of the obtained isomorphic knowledge graph is greatly improved. Therefore, the accuracy of the generated audit evaluation information is improved.

And 305, generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and the pre-trained quality evaluation model.

In some embodiments, the execution subject may generate audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the set of homogeneous knowledge graphs, and the pre-trained quality evaluation model. Wherein, the quality evaluation model comprises: the system comprises a feature extraction model set, a first classification model, a feature fusion model and a second classification model. For example, the feature extraction model in the feature extraction model set described above may be a VGG-16 model. The feature fusion model may be an LSTM model. The first classification network model and the second classification model may be convolutional neural networks with classification layers.

Wherein, the execution main body evaluates the model according to the target knowledge graph, the isomorphic knowledge graph set and the pre-trained quality, generating audit evaluation information corresponding to the target audit report text may include the following steps:

the first step is that the isomorphic knowledge maps in the target knowledge map and the isomorphic knowledge map set are input into the feature extraction models in the feature extraction model set in parallel to generate a first feature map feature information set and a second feature map feature information set.

The first feature map feature information is feature information of a target knowledge map extracted by the feature extraction model. The second map feature information is feature information of the isomorphic knowledge map extracted by the feature extraction model.

And inputting the first feature map feature information and the second feature map feature information set into the first classification model to generate a classification result.

And thirdly, inputting the first feature map feature information and at least one second map feature information corresponding to the classification result into the feature fusion model to generate third map feature information.

And fourthly, inputting the third map characteristic information into the second classification model to generate the audit evaluation information.

Step 306, sending a communication request to the communication terminal connected with the target display terminal.

In some embodiments, the execution subject may transmit a communication request to a communication terminal connected to the target display terminal. The communication terminal is in communication connection with the target display terminal. The communication terminal may be a terminal having an information transmitting/receiving function. The communication request may be a request for a communication connection with the communication terminal.

Step 307, in response to receiving the communication connection request sent by the communication terminal, an end-to-end communication link is created with the communication terminal.

In some embodiments, the execution body may create an end-to-end communication link with the communication terminal in response to receiving a communication connection request sent by the communication terminal. The communication connection request may be a request for requesting a communication connection with the execution subject.

And 308, in response to the determination that the communication link is successfully established, transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link.

In some embodiments, the execution subject may be to transmit, to the communication terminal via the communication link, encrypted text corresponding to the audit trail information in response to determining that the communication link was successfully created. The encrypted text corresponding to the audit evaluation information may be a text generated after being encrypted by an encryption algorithm. The encryption algorithm may be a symmetric encryption algorithm.

Optionally, the executing body may further execute the following steps:

first, an encryption key is constructed according to the communication address and the transmission time stamp of the communication terminal.

The communication Address may be an IP (Internet Protocol) Address of the communication terminal, or may be a Media Access Control (MAC) Address.

As an example, first, the execution body may perform character string concatenation on the communication address and the transmission time stamp to generate a concatenated character string. Then, the execution subject may hash the concatenated string by using a hash algorithm to generate the encryption key.

And secondly, performing information encryption on the audit evaluation information through the encryption secret key to generate the encrypted text.

The execution main body may use the encryption key as a private key, and encrypt the audit evaluation information by using a public key through an asymmetric encryption algorithm to generate the encrypted text.

And 309, sending the audit evaluation information to a target display terminal for displaying.

In some embodiments, the specific implementation of step 309 and the technical effect thereof may refer to step 206 in the embodiment corresponding to fig. 2, which is not described herein again.

As can be seen from fig. 3, compared to the description of some embodiments corresponding to fig. 2, the present disclosure first determines an isomorphic knowledge graph similar to the target knowledge graph by determining the entity node similarity value and the relationship edge similarity value, and by weighted summation, the speed of determining the isomorphic knowledge graph is faster. In addition, the target knowledge graph and the isomorphic knowledge graph are subjected to feature extraction through a plurality of parallel feature extraction models, and the obtained graph features are greatly enriched. Therefore, the accuracy of the generated audit evaluation information is improved,

with further reference to fig. 4, as an implementation of the methods illustrated in the above figures, the present disclosure provides some embodiments of a knowledge-graph-based audit trail information generation apparatus, which correspond to those method embodiments illustrated in fig. 2, and which may be particularly applicable in various electronic devices.

As shown in fig. 4, the knowledge-graph-based audit trail information generation apparatus 400 of some embodiments includes: an acquisition unit 401, a text washing unit 402, a first generation unit 403, a determination unit 404, a second generation unit 405, and a display unit 406. An obtaining unit configured to obtain a target audit report text; the text cleaning unit is configured to perform text cleaning on the target audit report text to generate text information; the first generating unit is configured to generate a target knowledge graph corresponding to the text information; the determining unit is configured to determine an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set; a second generating unit, configured to generate audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model; and the display unit is configured to send the audit evaluation information to a target display terminal for displaying.

In some optional implementations of some embodiments, the text washing unit 402 is further configured to: reading the target audit report text to generate candidate text information; performing entity identification on the candidate text information to generate a first entity information set, wherein first entity information in the first entity information set includes: entity location information and entity information; determining entity similarity between the entity information included in the first entity information and each entity information in the target entity library to generate a similarity value; screening entity information of which the corresponding similarity value is located in a target interval from the target entity library, and taking the entity information as candidate entity information to obtain a candidate entity information set; performing entity replacement on the first entity information included in the candidate text information according to entity position information and target candidate entity information included in the first entity information, wherein the target candidate entity information is candidate entity information meeting a first screening condition in the candidate entity information set; and determining the candidate text information after the entity replacement as the text information.

In some optional implementations of some embodiments, the first generating unit 403 is further configured to: carrying out entity identification on the text information to generate a second entity information set; entity removing is carried out on second entity information in the second entity information set to generate removed entity information, and a removed entity information set is obtained; determining the relationship information corresponding to each entity information in the entity information set after the elimination to obtain a relationship information group set; and generating the target knowledge graph according to the entity information set after the elimination and the relation information group set.

In some optional implementations of some embodiments, the determining unit 404 is further configured to: screening a knowledge graph meeting second screening conditions from a knowledge graph library to serve as a candidate knowledge graph to obtain a candidate knowledge graph set, wherein the second screening conditions are that the number of entity nodes contained in the knowledge graph is greater than or equal to the number of entity nodes contained in the target knowledge graph, and the number of relation edges contained in the knowledge graph is greater than or equal to the number of relation edges contained in the target knowledge graph; executing the following processing steps on each entity node in the target knowledge graph: constructing a feature vector of the entity node to generate a first entity node feature vector; constructing a feature vector for at least one relationship edge connected with the entity node to generate a first relationship edge feature vector; for each candidate knowledge-graph in the set of candidate knowledge-graphs, performing the following processing steps: constructing a feature vector for each entity node in the candidate knowledge graph to generate a second entity node feature vector to obtain a second entity node feature vector sequence; constructing a feature vector for each target relationship edge group in the candidate knowledge graph to generate a second relationship edge feature vector sequence, wherein the target relationship edge group is at least one relationship edge connected with the entity nodes of the candidate knowledge graph; determining the feature similarity of each first entity node feature vector in the obtained first entity node feature vector sequence and each second entity node feature vector in the second entity node feature vector sequence to obtain an entity node similarity value group sequence; determining the feature similarity of each first relation edge feature vector in the obtained first relation edge feature vector sequence and each second relation edge feature vector in the second relation edge feature vector sequence to obtain a relation edge similarity value group sequence; screening out the entity node similarity value with the maximum value from each entity node similarity value group in the entity node similarity value group sequence as a first candidate value to obtain a first candidate value set; screening out a relation side similarity value with the largest value from each relation side similarity value group in the relation side similarity value group sequence, and taking the relation side similarity value as a second candidate value to obtain a second candidate value group; carrying out weighted summation on a first candidate value in the first candidate value set and a second candidate value in the second candidate value set so as to generate a map similarity numerical value corresponding to the candidate knowledge map; and screening out the candidate knowledge graphs with the corresponding graph similarity numerical values meeting the third screening condition from the candidate knowledge graph set to serve as isomorphic knowledge graphs, and obtaining the isomorphic knowledge graph set. In practical situations, knowledge maps corresponding to different audit report texts often contain a part of different feature information. Therefore, by determining the isomorphic knowledge graph set corresponding to the target knowledge graph, the richness of the characteristics can be increased, and the generated audit evaluation information can be more accurate. Therefore, how to determine the isomorphic knowledge graph set corresponding to the target knowledge graph is particularly important. First, the isomorphic knowledge graph should contain the number of similar relationship edges and the number of entity nodes. Therefore, the present disclosure coarsely screens the knowledge-graph through the second screening condition to filter out useless knowledge-graphs. In addition, the graph structures of the isomorphic knowledge graphs should be similar, so the method and the device for determining the isomorphic knowledge graphs determine the isomorphic knowledge graphs in a mode that feature vectors are constructed on entity nodes and relation edges connected with the entity nodes, and similarity is calculated according to the constructed feature vectors. By the method, the accuracy of the obtained isomorphic knowledge graph is greatly improved. Therefore, the accuracy of the generated audit evaluation information is improved.

In some optional implementations of some embodiments, the quality assessment model includes: the method comprises the steps of collecting a feature extraction model, a first classification model, a feature fusion model and a second classification model; and the second generating unit 405 is further configured to: inputting the isomorphic knowledge maps in the target knowledge map and the isomorphic knowledge map set into the feature extraction models in the feature extraction model set in parallel to generate a first feature map feature information set and a second feature map feature information set; inputting the first feature map feature information and the second feature map feature information set into the first classification model to generate a classification result; inputting the first feature map feature information and at least one second map feature information corresponding to the classification result into the feature fusion model to generate third map feature information; and inputting the third map feature information into the second classification model to generate the audit evaluation information.

In some optional implementations of some embodiments, before sending the audit trail information to a target display terminal for display, the apparatus 400 further includes: sending a communication request to a communication terminal connected with the target display terminal; in response to receiving a communication connection request sent by the communication terminal, establishing an end-to-end communication link with the communication terminal; and in response to the fact that the communication link is successfully established, transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link.

In some optional implementation manners of some embodiments, before the transmitting, to the communication terminal, the encrypted text corresponding to the audit trail information through the communication link, the apparatus 400 further includes: constructing an encryption key according to the communication address and the sending time stamp of the communication terminal; and encrypting the audit evaluation information through the encryption key to generate the encrypted text.

It will be understood that the elements described in the apparatus 400 correspond to various steps in the method described with reference to fig. 2. Thus, the operations, features and advantages described above with respect to the method are also applicable to the apparatus 400 and the units included therein, and are not described herein again.

Referring now to FIG. 5, a block diagram of an electronic device (such as computing device 101 shown in FIG. 1) 500 suitable for use in implementing some embodiments of the present disclosure is shown. The electronic device shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.

As shown in fig. 5, electronic device 500 may include a processing means (e.g., central processing unit, graphics processor, etc.) 501 that may perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM) 502 or a program loaded from a storage means 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data necessary for the operation of the electronic apparatus 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.

Generally, the following devices may be connected to the I/O interface 505: input devices 506 including, for example, a touch screen, touch pad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; output devices 507 including, for example, a Liquid Crystal Display (LCD), speakers, vibrators, and the like; storage devices 508 including, for example, magnetic tape, hard disk, etc.; and a communication device 509. The communication means 509 may allow the electronic device 500 to communicate with other devices wirelessly or by wire to exchange data. While fig. 5 illustrates an electronic device 500 having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided. Each block shown in fig. 5 may represent one device or may represent multiple devices as desired.

In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flow diagrams may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In some such embodiments, the computer program may be downloaded and installed from a network via the communication means 509, or installed from the storage means 508, or installed from the ROM 502. The computer program, when executed by the processing device 501, performs the above-described functions defined in the methods of some embodiments of the present disclosure.

It should be noted that the computer readable medium described in some embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.

In some embodiments, the clients, servers may communicate using any currently known or future developed network Protocol, such as HTTP (HyperText Transfer Protocol), and may interconnect with any form or medium of digital data communication (e.g., a communications network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the Internet (e.g., the Internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed network.

The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring a target audit report text; performing text cleaning on the target audit report text to generate text information; generating a target knowledge graph corresponding to the text information; determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set; generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model; and sending the audit evaluation information to a target display terminal for display.

Computer program code for carrying out operations for embodiments of the present disclosure may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider).

The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in some embodiments of the present disclosure may be implemented by software, and may also be implemented by hardware. The described units may also be provided in a processor, and may be described as: a processor includes an acquisition unit, a text cleansing unit, a first generation unit, a determination unit, a second generation unit, and a display unit. Where the names of these elements do not in some cases constitute a limitation on the elements themselves, for example, an acquisition element may also be described as an "element that acquires target audit report text".

The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems on a chip (SOCs), complex Programmable Logic Devices (CPLDs), and the like.

The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above-mentioned features, but also encompasses other embodiments in which any combination of the above-mentioned features or their equivalents is made without departing from the inventive concept as defined above. For example, the above features and (but not limited to) technical features with similar functions disclosed in the embodiments of the present disclosure are mutually replaced to form the technical solution.

Claims (11)

1. A knowledge graph-based audit evaluation information generation method comprises the following steps:

acquiring a target audit report text;

performing text cleaning on the target audit report text to generate text information;

generating a target knowledge graph corresponding to the text information;

determining an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set;

generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model;

and sending the audit evaluation information to a target display terminal for display.

2. The method of claim 1, wherein the text-washing the target audit report text to generate textual information comprises:

reading the target audit report text to generate candidate text information;

performing entity identification on the candidate text information to generate a first entity information set, wherein first entity information in the first entity information set comprises: entity location information and entity information;

determining entity similarity between entity information included in the first entity information and each entity information in a target entity library to generate a similarity value;

screening out entity information of which the corresponding similarity value is located in a target interval from the target entity library, and taking the entity information as candidate entity information to obtain a candidate entity information set;

performing entity replacement on the first entity information included in the candidate text information according to entity position information and target candidate entity information included in the first entity information, wherein the target candidate entity information is candidate entity information meeting a first screening condition in the candidate entity information set;

and determining the candidate text information after entity replacement as the text information.

3. The method of claim 1, wherein the generating a target knowledge-graph corresponding to the textual information comprises:

performing entity identification on the text information to generate a second entity information set;

entity removing is carried out on second entity information in the second entity information set to generate removed entity information, and a removed entity information set is obtained;

determining the corresponding relation information of each entity information in the entity information set after being removed to obtain a relation information group set;

and generating the target knowledge graph according to the entity information set after the elimination and the relation information group set.

4. The method of claim 3, wherein the determining the homogeneous knowledge-graph corresponding to the target knowledge-graph to obtain a homogeneous knowledge-graph set comprises:

screening a knowledge graph which meets second screening conditions from a knowledge graph library to serve as a candidate knowledge graph to obtain a candidate knowledge graph set, wherein the second screening conditions are that the number of entity nodes contained in the knowledge graph is greater than or equal to the number of entity nodes contained in the target knowledge graph, and the number of relation edges contained in the knowledge graph is greater than or equal to the number of relation edges contained in the target knowledge graph;

performing the following processing steps for each entity node in the target knowledge graph:

constructing a feature vector for the entity node to generate a first entity node feature vector;

constructing a feature vector for at least one relationship edge connected with the entity node to generate a first relationship edge feature vector;

for each candidate knowledge-graph of the set of candidate knowledge-graphs, performing the following processing steps:

constructing a feature vector for each entity node in the candidate knowledge graph to generate a second entity node feature vector to obtain a second entity node feature vector sequence;

constructing a feature vector for each target relationship edge group in the candidate knowledge graph to generate a second relationship edge feature vector sequence, wherein the target relationship edge group is at least one relationship edge connected with the entity nodes of the candidate knowledge graph;

determining the feature similarity of each first entity node feature vector in the obtained first entity node feature vector sequence and each second entity node feature vector in the second entity node feature vector sequence to obtain an entity node similarity value group sequence;

determining the feature similarity of each first relation edge feature vector in the obtained first relation edge feature vector sequence and each second relation edge feature vector in the second relation edge feature vector sequence to obtain a relation edge similarity value group sequence;

screening out the entity node similarity value with the maximum value from each entity node similarity value group in the entity node similarity value group sequence as a first candidate value to obtain a first candidate value set;

screening out a relation side similarity value with the largest value from each relation side similarity value group in the relation side similarity value group sequence, and taking the relation side similarity value as a second candidate value to obtain a second candidate value group;

performing weighted summation on a first candidate value in the first candidate value set and a second candidate value in the second candidate value set to generate a map similarity numerical value corresponding to the candidate knowledge map;

and screening out a corresponding map similarity numerical value which meets a third screening condition from the candidate knowledge map set to serve as an isomorphic knowledge map, so as to obtain the isomorphic knowledge map set.

5. The method of claim 4, wherein the quality assessment model comprises: the method comprises the steps of collecting a feature extraction model, a first classification model, a feature fusion model and a second classification model; and

generating audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model, wherein the audit evaluation information comprises:

inputting the target knowledge graph and the isomorphic knowledge graph in the isomorphic knowledge graph set into a feature extraction model in the feature extraction model set in parallel to generate a first feature graph feature information set and a second feature graph feature information set;

inputting the first feature profile feature information and the second profile feature information set into the first classification model to generate a classification result;

inputting the first feature map feature information and at least one second map feature information corresponding to the classification result into the feature fusion model to generate third map feature information;

inputting the third graph feature information into the second classification model to generate the audit trail information.

6. The method of claim 1, wherein prior to said sending the audit trail information to a target display terminal for display, the method further comprises:

sending a communication request to a communication terminal connected with the target display terminal;

in response to receiving a communication connection request sent by the communication terminal, establishing an end-to-end communication link with the communication terminal;

and in response to the fact that the communication link is successfully established, transmitting the encrypted text corresponding to the audit evaluation information to the communication terminal through the communication link.

7. The method of claim 6, wherein prior to said transmitting encrypted text corresponding to said audit trail information to said communication terminal over said communication link, said method further comprises:

constructing an encryption key according to the communication address and the sending timestamp of the communication terminal;

and encrypting the audit evaluation information through the encryption key to generate the encrypted text.

8. A knowledge-graph-based audit assessment information generation device comprises:

an obtaining unit configured to obtain a target audit report text;

a text cleaning unit configured to perform text cleaning on the target audit report text to generate text information;

a first generating unit configured to generate a target knowledge graph corresponding to the text information;

the determining unit is configured to determine an isomorphic knowledge graph corresponding to the target knowledge graph to obtain an isomorphic knowledge graph set;

a second generation unit configured to generate audit evaluation information corresponding to the target audit report text according to the target knowledge graph, the isomorphic knowledge graph set and a pre-trained quality evaluation model;

a display unit configured to send the audit assessment information to a target display terminal for display.

9. The knowledge-graph-based audit trail assessment information generation apparatus according to claim 8, wherein the determination unit is further configured to:

screening a knowledge graph meeting second screening conditions from a knowledge graph library to serve as a candidate knowledge graph to obtain a candidate knowledge graph set, wherein the second screening conditions are that the number of entity nodes contained in the knowledge graph is greater than or equal to the number of entity nodes contained in the target knowledge graph, and the number of relation edges contained in the knowledge graph is greater than or equal to the number of relation edges contained in the target knowledge graph;

performing the following processing steps for each entity node in the target knowledge graph:

constructing a feature vector for the entity node to generate a first entity node feature vector;

constructing a feature vector for at least one relationship edge connected with the entity node to generate a first relationship edge feature vector;

for each candidate knowledge-graph of the set of candidate knowledge-graphs, performing the following processing steps:

constructing a feature vector for each entity node in the candidate knowledge graph to generate a second entity node feature vector to obtain a second entity node feature vector sequence;

constructing a feature vector for each target relationship edge group in the candidate knowledge graph to generate a second relationship edge feature vector sequence, wherein the target relationship edge group is at least one relationship edge connected with the entity nodes of the candidate knowledge graph;

determining the feature similarity of each first entity node feature vector in the obtained first entity node feature vector sequence and each second entity node feature vector in the second entity node feature vector sequence to obtain an entity node similarity value group sequence;

determining the feature similarity of each first relation edge feature vector in the obtained first relation edge feature vector sequence and each second relation edge feature vector in the second relation edge feature vector sequence to obtain a relation edge similarity value group sequence;

screening out the entity node similarity value with the maximum value from each entity node similarity value group in the entity node similarity value group sequence as a first candidate value to obtain a first candidate value set;

screening out a relation side similarity value with the largest value from each relation side similarity value group in the relation side similarity value group sequence, and taking the relation side similarity value as a second candidate value to obtain a second candidate value group;

performing weighted summation on a first candidate value in the first candidate value set and a second candidate value in the second candidate value set to generate a map similarity numerical value corresponding to the candidate knowledge map;

and screening out a corresponding map similarity numerical value which meets a third screening condition from the candidate knowledge map set to serve as an isomorphic knowledge map, so as to obtain the isomorphic knowledge map set.

10. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon;

when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-7.

11. A computer-readable medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 7.

Images