Municipal engineering supervision data standardization management method

Abstract

The invention relates to the field of intelligent management, in particular to a municipal engineering supervision data standardization management method. According to the method, a natural language processing technology based on deep learning is adopted to perform abstract generation on municipal engineering supervision data to be managed, and then fine-grained semantic analysis and reverse attention selection are performed on text abstracts of the supervision data, so that significance information in the supervision data is mined; therefore, the type identification and automatic filing processing of the municipal engineering supervision data are realized. Therefore, the efficiency and accuracy of data management can be effectively improved, standardized and normalized management of supervision data is promoted, and powerful support is provided for municipal engineering construction.

Description

Technical Field

[0001] The present application relates to the field of intelligent management, and specifically to a method for standardized management of municipal engineering supervision data. Background Art

[0002] Municipal engineering supervision data refers to important documents and materials that are generated or obtained by engineering supervision units and project supervision agencies in the process of fulfilling construction engineering supervision contracts (construction engineering supervision and related services), reflecting the objective conditions of project construction and the performance of supervision duties. In the field of municipal engineering construction, the management of supervision data is an important link to ensure project quality, safety, progress and cost control.

[0003] With the continuous advancement of urban infrastructure construction, the amount of supervision data has increased dramatically, and the complexity and diversity of its content have also put forward higher requirements for data management. The traditional supervision data management method relies on manual review to organize and archive, which is not only time-consuming and labor-intensive and inefficient, but also prone to errors and omissions, and it is difficult to meet the needs of current large-scale municipal engineering construction.

[0004] Therefore, an intelligent standardized management method for municipal engineering supervision data is expected. Summary of the invention

[0005] This application is made in view of the above problems. One purpose of this application is to provide a standardized management method for municipal engineering supervision data.

[0006] The embodiment of the present application provides a method for standardizing the management of municipal engineering supervision data, which is characterized by comprising:

[0007] Obtain supervision information of municipal engineering projects to be managed;

[0008] Generate a summary of the municipal engineering supervision data to be managed to obtain a text summary of the municipal engineering supervision data;

[0009] Performing segment-granularity-based semantic coding on the municipal engineering supervision document text summary to obtain a sequence of segment description semantic coding feature vectors of the municipal engineering supervision document text summary;

[0010] Performing feature reverse attention selection on the sequence of semantic coding feature vectors describing the text summary segment of the municipal engineering supervision data to obtain a sequence of reverse attention selection market engineering supervision data segment granularity semantic feature vectors;

[0011] Determine the classification result of the data type based on the sequence of the granularity semantic feature vectors of the reverse attention selection market engineering supervision data segment;

[0012] Based on the classification result, the municipal engineering supervision data to be managed are archived.

[0013] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the summary of the municipal engineering supervision data to be managed is generated to obtain a text summary of the municipal engineering supervision data, including:

[0014] The municipal engineering supervision data to be managed is input into a text summary generator based on the AIGC model to obtain a text summary of the municipal engineering supervision data.

[0015] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the municipal engineering supervision data text summary is semantically encoded based on segment granularity to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description, including:

[0016] Segment processing is performed on the municipal engineering supervision document text summary to obtain a sequence of municipal engineering supervision document text summary segment descriptions;

[0017] Semantic coding is performed on each municipal engineering supervision document text summary segment description in the sequence of the municipal engineering supervision document text summary segment description to obtain a sequence of semantic coding feature vectors of the municipal engineering supervision document text summary segment description.

[0018] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, each municipal engineering supervision data text summary segment description in the sequence of municipal engineering supervision data text summary segment description is semantically encoded to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description, including:

[0019] Each municipal engineering supervision document text summary segment description in the sequence of the municipal engineering supervision document text summary segment description is input into the supervision document semantic understander based on the Bert model to obtain a sequence of the municipal engineering supervision document text summary segment description semantic encoding feature vectors.

[0020] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the sequence of semantic encoding feature vectors describing the text summary segment of the municipal engineering supervision data is subjected to feature reverse attention selection to obtain a sequence of reverse attention selection market engineering supervision data segment granularity semantic feature vectors, including:

[0021] Extracting significant features from the sequence of semantically encoded feature vectors describing the text summary segment of the municipal engineering supervision data to obtain a significant information representation vector of the municipal engineering supervision data segment;

[0022] Taking the significant information representation vector of the market engineering supervision data segment as the self-supervised cluster center representation vector, the sequence of semantic encoding feature vectors describing the text summary segment of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain a sequence of reverse attention selected granular semantic feature vectors of the market engineering supervision data segment.

[0023] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the sequence of semantic coding feature vectors describing the text summary segment of the municipal engineering supervision data is subjected to significant feature extraction to obtain a significant information representation vector of the market engineering supervision data segment, including:

[0024] The maximum values ​​of the semantic coding feature vectors of each municipal engineering supervision data text summary segment description in the sequence of the semantic coding feature vectors of the municipal engineering supervision data text summary segment description are extracted to obtain the municipal engineering supervision data segment significant information representation vector composed of multiple maximum values.

[0025] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the significant information representation vector of the market engineering supervision data segment is used as the self-supervised cluster center representation vector, and the sequence of the semantic encoding feature vectors of the text summary segment description of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain the sequence of the reverse attention selected market engineering supervision data segment granular semantic feature vectors, including:

[0026] Calculating the anti-correlation coefficient between each municipal engineering supervision document text summary segment description semantic coding feature vector in the sequence of the municipal engineering supervision document text summary segment description semantic coding feature vector and the self-supervised cluster center representation vector to obtain a sequence of anti-correlation coefficients;

[0027] Inputting the sequence of anti-correlation coefficients into a Sigmoid activation function to obtain a sequence of normalized anti-correlation weight coefficients;

[0028] Using each normalized anti-correlation weight coefficient in the sequence of normalized anti-correlation weight coefficients as a weight, weighting each municipal engineering supervision material text summary segment description semantic coding feature vector in the sequence of municipal engineering supervision material text summary segment description semantic coding feature vectors respectively to obtain a sequence of anti-reinforcement market engineering supervision material segment granularity semantic feature vectors;

[0029] Calculate the positional difference between each group of corresponding municipal engineering supervision data text summary segment description semantic coding feature vectors and anti-reinforcement market engineering supervision data segment granularity semantic feature vectors in the sequence of the municipal engineering supervision data text summary segment description semantic coding feature vectors and the sequence of the anti-reinforcement market engineering supervision data segment granularity semantic feature vectors to obtain the sequence of the reverse attention selection market engineering supervision data segment granularity semantic feature vectors.

[0030] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the anti-correlation coefficient between each municipal engineering supervision data text summary segment description semantic coding feature vector in the sequence of the municipal engineering supervision data text summary segment description semantic coding feature vector and the self-supervised cluster center representation vector is calculated to obtain a sequence of anti-correlation coefficients, including:

[0031] Calculate the covariance matrix between the semantic encoding feature vector of the municipal engineering supervision material text summary segment description and the self-supervised cluster center representation vector;

[0032] Calculate the position-by-position subtraction between the semantic encoding feature vector of the municipal engineering supervision material text summary segment description and the self-supervised cluster center representation vector to obtain a semantic differential feature vector;

[0033] The anti-correlation coefficient is obtained by calculating the square root of a product result of a transposed vector of the semantic differential feature vector, the covariance matrix, and the semantic differential feature vector.

[0034] For example, according to the standardized management method of municipal engineering supervision data of an embodiment of the present application, the classification result of the data type is determined based on the sequence of the segment granularity semantic feature vectors of the reverse attention selection market engineering supervision data, including:

[0035] After the sequence of reverse attention-selected market engineering supervision data segment-granularity semantic feature vectors is cascaded into a market engineering supervision data text summary semantic salient coding vector, the market engineering supervision data text summary semantic salient coding vector is input into a classifier-based data classifier to obtain the classification result, which is the data type label of the municipal engineering supervision data to be managed.

[0036] According to the standardized management method of municipal engineering supervision data in the embodiment of the present application, it uses natural language processing technology based on deep learning to generate summaries of the municipal engineering supervision data to be managed, and then conducts fine-grained semantic analysis and reverse attention selection on the text summaries of the supervision data to mine the significant information in the supervision data, thereby realizing type recognition and automatic filing processing of the municipal engineering supervision data. In this way, the efficiency and accuracy of data management can be effectively improved, the standardized and normalized management of supervision data can be promoted, and strong support can be provided for municipal engineering construction. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0038] Figure 1 A schematic diagram of the application architecture of the municipal engineering supervision data standardization management method in the embodiment of the present application is shown;

[0039] Figure 2 A flow chart of a method for standardizing management of municipal engineering supervision data in an embodiment of the present application is shown;

[0040] Figure 3 A flowchart of sub-step S530 of the method for standardizing management of municipal engineering supervision data in an embodiment of the present application is shown;

[0041] Figure 4 A flowchart of sub-step S540 of the method for standardizing management of municipal engineering supervision data in an embodiment of the present application is shown;

[0042] Figure 5 A schematic diagram of the structure of a standardized management system for municipal engineering supervision data in an embodiment of the present application is shown; and

[0043] Figure 6 An application scenario diagram of the standardized management method for municipal engineering supervision data in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0045] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intentions of those skilled in the art, precedents, or new technologies in the art. In addition, specific terms may be selected, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.

[0046] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0047] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0048] Figure 1 A schematic diagram of the application architecture of the standardized management method for municipal engineering supervision data in an embodiment of the present application is shown, including a server 100 and a terminal device 200.

[0049] The terminal device 200 and the server 100 can be connected via the Internet to achieve mutual communication. Optionally, the above-mentioned Internet uses standard communication technology and / or protocol. The Internet is usually the Internet, but it can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a dedicated network or a virtual private network. In some embodiments, the data exchanged through the network is represented by technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec) can also be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.

[0050] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, the server 100 may include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130, and an output device 140, etc. The input device 130 may include a keyboard, a mouse, a touch screen, etc., and the output device 140 may include a display device, such as a liquid crystal display (Liquid Crystal Display, LCD), a cathode ray tube (Cathode Ray Tube, CRT), etc.

[0051] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In the embodiment of the present application, the memory 120 may be used to store the program of the standardized management method for municipal engineering supervision data in the embodiment of the present application.

[0052] The processor 110 calls the program instructions stored in the memory 120, and the processor 110 is used to execute the steps of any one of the municipal engineering supervision data standardization management methods in the embodiments of the present application according to the obtained program instructions.

[0053] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.

[0054] The following is a non-restrictive explanation of the standardized management method of municipal engineering supervision data provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present application.

[0055] In response to the above technical problems, the technical concept of this application is to use natural language processing technology based on deep learning to generate summaries of municipal engineering supervision materials to be managed, and then conduct fine-grained semantic analysis and reverse attention selection on the text summaries of the supervision materials to mine the significant information in the supervision materials, thereby realizing the type recognition and automatic filing of municipal engineering supervision materials. In this way, the efficiency and accuracy of data management can be effectively improved, the standardization and normalization of supervision materials can be promoted, and strong support can be provided for municipal engineering construction.

[0056] Based on this, Figure 2 The flowchart of the municipal engineering supervision data standardization management method in the embodiment of the present application is shown. For example, the municipal engineering supervision data standardization management method can be executed by a server, which can be Figure 1 The server 100 shown in FIG. Figure 2As shown, according to the standardized management method of municipal engineering supervision data of the embodiment of the present application, the steps include: S510, obtaining the municipal engineering supervision data to be managed; S520, summarizing the municipal engineering supervision data to be managed to obtain a text summary of the municipal engineering supervision data; S530, semantically encoding the text summary of the municipal engineering supervision data based on segment granularity to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description; S540, performing feature reverse attention selection on the sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description to obtain a sequence of reverse attention selected market engineering supervision data segment granularity semantic feature vectors; S550, determining the classification result of the data type based on the sequence of reverse attention selected market engineering supervision data segment granularity semantic feature vectors; S560, archiving the municipal engineering supervision data to be managed based on the classification result.

[0057] Specifically, in the technical solution of the present application, first, the municipal engineering supervision data to be managed are obtained. Municipal engineering supervision data refers to the documents and materials that are formed or obtained by the engineering supervision unit and the project supervision agency in the process of performing the construction engineering supervision contract (construction engineering supervision and related services), reflecting the objective situation of the project construction, in a certain form of written materials (including text and images, etc.), records, and organized and preserved. Municipal engineering supervision data is an important manifestation of the quality work of the project supervision agency during the construction process, a necessary condition for the completion acceptance of the project quality, an important part of the urban construction archives, and an important written basis for verifying the performance of the supervision contract. Therefore, the compilation / collection / daily management and preservation of municipal engineering supervision data, until the classification management, filing, binding, delivery to the relevant departments or archiving after completion, should be subject to standardized and standardized management, and a corresponding management system inspection and acceptance system should be established.

[0058] The information and documents generated in the process of project supervision can be divided into the following categories according to the attributes of the documents: 1. Preparation materials: including the supervision plan, supervision implementation rules, supervision witness plan, supervision monthly report, engineering quality assessment report and supervision work summary prepared by the project supervision agency; 2. Issuance materials: including the appointment document for the establishment of the project supervision agency issued by the supervision company, various supervision instructions and supervision reports, work contact forms and other documents issued by the project supervision agency; 3. Review and verification materials: including the various review and verification forms and their attachments submitted by the construction unit for review and verification by the project supervision agency; 4. Record materials: including the supervision log, meeting minutes, on-site records, witness records, parallel inspection records and other documents formed by the project supervision agency in the supervision process; 5. Ledger materials: including the work ledger established by the project supervision agency for the acceptance of materials, components and equipment, and witness sampling and inspection in combination with the specific circumstances of the supervision project.

[0059] Next, considering that municipal engineering supervision data often contain a large amount of detailed descriptions of engineering supervision information, directly processing them is not only computationally expensive, but also inefficient. Therefore, this application uses a text summary generator based on the AIGC model to perform semantic parsing and summary generation on the municipal engineering supervision data to be managed, so as to use the powerful generation ability of the AIGC model to compress long text information into concise and refined text summaries of municipal engineering supervision data, thereby retaining key information and removing redundant content, so as to facilitate subsequent data analysis and processing.

[0060] Correspondingly, in step S520, a summary of the municipal engineering supervision data to be managed is generated to obtain a text summary of the municipal engineering supervision data, including: inputting the municipal engineering supervision data to be managed into a text summary generator based on the AIGC model to obtain the text summary of the municipal engineering supervision data.

[0061] Secondly, considering that the text summary of the municipal engineering supervision materials may contain multiple different supervision information topics, in order to more accurately understand the information content of the municipal engineering supervision materials, the text summary of the municipal engineering supervision materials is further segmented to form a sequence of municipal engineering supervision materials text summary segment descriptions, so as to facilitate independent semantic analysis of each topic paragraph.

[0062] Furthermore, each of the municipal engineering supervision document text summary segment descriptions in the sequence of the municipal engineering supervision document text summary segment descriptions is semantically encoded to extract the content semantic features of each text summary segment description, and convert it from an unstructured text form into a computer-understandable mathematical representation to generate a sequence of semantically encoded feature vectors of the municipal engineering supervision document text summary segment descriptions. In a specific example of the present application, a supervision document semantic understander based on the Bert model is used to implement this semantic encoding process. The Bert model is based on a bidirectional Transformer architecture, which can capture the content meaning and bidirectional contextual relationship of each word unit in the text, so as to more accurately understand the semantic content of each municipal engineering supervision document text summary segment description.

[0063] Accordingly, in step S530, if Figure 3As shown, the municipal engineering supervision data text summary is semantically encoded based on segment granularity to obtain a sequence of semantic encoding feature vectors of municipal engineering supervision data text summary segment descriptions, including: S531, segmenting the municipal engineering supervision data text summary to obtain a sequence of municipal engineering supervision data text summary segment descriptions; S532, semantically encoding each municipal engineering supervision data text summary segment description in the sequence of municipal engineering supervision data text summary segment descriptions to obtain a sequence of semantic encoding feature vectors of municipal engineering supervision data text summary segment descriptions.

[0064] Among them, in step S532, each municipal engineering supervision data text summary segment description in the sequence of municipal engineering supervision data text summary segment description is semantically encoded to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description, including: each municipal engineering supervision data text summary segment description in the sequence of municipal engineering supervision data text summary segment description is input into a supervision data semantic understander based on the Bert model to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description.

[0065] Then, in order to further improve the feature representation accuracy and discrimination of the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data, and reduce the information redundancy and noise interference therein. In the technical scheme of the present application, the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data is further subjected to feature enhancement processing. Specifically, first, the maximum value of each semantic coding feature vector of the text summary segment description of the municipal engineering supervision data in the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data is extracted to obtain the significant information feature representation in each summary segment description, and obtain a more refined and representative market engineering supervision data segment significant information representation vector, which provides an important reference benchmark for the feature enhancement processing of the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data.

[0066] Next, the significant information representation vector of the market engineering supervision data segment is used as the self-supervised cluster center representation vector, and the sequence of the semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain a sequence of reverse attention selected granular semantic feature vectors of the market engineering supervision data segment. That is, based on the key information of the supervision data contained in the significant information representation vector of the market engineering supervision data segment, the model is guided to learn more distinguishing features in the sequence of the semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data. Specifically, first, the anti-correlation coefficient between the semantic coding feature vectors of each municipal engineering supervision data text summary segment description and the significant information representation vector of the market engineering supervision data segment is calculated to quantify the heterogeneity between the two, that is, the degree of semantic irrelevance between the two. Then, the anti-correlation coefficient is converted into a standardized weight coefficient through normalization processing, so as to weight and modulate the original feature distribution of the sequence of the semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data, strengthen the features that are greatly different from the cluster center, and suppress the features that are similar to the cluster center to enhance the distinguishability of the features. Then, by calculating the positional difference between the original feature distribution and the weighted enhanced feature distribution, irrelevant information is filtered out, and the feature components that are positively correlated with the cluster center are reversely selected, thereby significantly improving the expressive power of the features and the generalization performance of the model.

[0067] Accordingly, in step S540, if Figure 4 As shown, the sequence of semantic coding feature vectors of the description of the text summary segment of the municipal engineering supervision data is subjected to feature reverse attention selection to obtain a sequence of reverse attention selected market engineering supervision data segment granularity semantic feature vectors, including: S541, performing significant feature extraction on the sequence of semantic coding feature vectors of the description of the text summary segment of the municipal engineering supervision data to obtain a significant information representation vector of the market engineering supervision data segment; S542, using the significant information representation vector of the market engineering supervision data segment as the self-supervised cluster center representation vector, performing feature reverse attention selection based on the self-supervised cluster center on the sequence of semantic coding feature vectors of the description of the text summary segment of the municipal engineering supervision data to obtain a sequence of reverse attention selected market engineering supervision data segment granularity semantic feature vectors.

[0068] Among them, in step S541, significant features are extracted from the sequence of semantic coding feature vectors of the description of the text summary segment of the municipal engineering supervision data to obtain a significant information representation vector of the market engineering supervision data segment, including: extracting the maximum value of each semantic coding feature vector of the description of the text summary segment of the municipal engineering supervision data in the sequence of semantic coding feature vectors of the description of the text summary segment of the municipal engineering supervision data to obtain a significant information representation vector of the market engineering supervision data segment composed of multiple maximum values.

[0069] Wherein, in step S542, the significant information representation vector of the market engineering supervision data segment is used as the self-supervised cluster center representation vector, and the sequence of the semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain the sequence of the reverse attention selected market engineering supervision data segment granular semantic feature vectors, including: calculating the anti-correlation coefficients between each municipal engineering supervision data text summary segment description semantic coding feature vector in the sequence of the municipal engineering supervision data text summary segment description semantic coding feature vector and the self-supervised cluster center representation vector to obtain a sequence of anti-correlation coefficients; inputting the sequence of anti-correlation coefficients into the Sigmoid activation function to obtain a sequence of normalized anti-correlation weight coefficients; using the normalized Each normalized anti-correlation weight coefficient in the sequence of normalized anti-correlation weight coefficients is used as a weight, and each municipal engineering supervision data text summary segment description semantic coding feature vector in the sequence of municipal engineering supervision data text summary segment description semantic coding feature vectors is weighted respectively to obtain a sequence of anti-reinforcement market engineering supervision data segment granularity semantic feature vectors; the positional difference between each group of corresponding municipal engineering supervision data text summary segment description semantic coding feature vectors and anti-reinforcement market engineering supervision data segment granularity semantic feature vectors in the sequence of municipal engineering supervision data text summary segment description semantic coding feature vectors and the sequence of anti-reinforcement market engineering supervision data segment granularity semantic feature vectors is calculated to obtain the sequence of reverse attention selection market engineering supervision data segment granularity semantic feature vectors.

[0070] Specifically, the anti-correlation coefficients between each semantic coding feature vector of the text summary segment description of the municipal engineering supervision material and the self-supervised cluster center representation vector in the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision material are calculated to obtain a sequence of anti-correlation coefficients, including: calculating the covariance matrix between the semantic coding feature vector of the text summary segment description of the municipal engineering supervision material and the self-supervised cluster center representation vector; calculating the position-by-position subtraction between the semantic coding feature vector of the text summary segment description of the municipal engineering supervision material and the self-supervised cluster center representation vector to obtain a semantic differential feature vector; calculating the square root of the product result of the transposed vector of the semantic differential feature vector, the covariance matrix and the semantic differential feature vector to obtain the anti-correlation coefficient.

[0071] In a specific example, the significant information representation vector of the market engineering supervision data segment is used as the self-supervised cluster center representation vector, and the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain the sequence of granular semantic feature vectors of the reverse attention selected market engineering supervision data segment, including: processing the sequence of semantic coding feature vectors of the text summary segment description of the municipal engineering supervision data with the following reverse attention selection formula to obtain the sequence of granular semantic feature vectors of the reverse attention selected market engineering supervision data segment, wherein the reverse attention selection formula is:

[0072]

[0073] s ai =sigmoid[s i (x i , v c )]

[0074] X f ={s a1 x1,s a2 x2, ..., s an ·x n}

[0075]

[0076] Among them, x1, x2, x i and x n They respectively represent the first, second, i-th and n-th municipal engineering supervision document text summary segment description semantic coding feature vectors in the sequence of the municipal engineering supervision document text summary segment description semantic coding feature vectors, and the value of n is the number of feature vectors in the sequence of the municipal engineering supervision document text summary segment description semantic coding feature vectors, v c represents the self-supervised cluster center representation vector, S represents the covariance matrix between the semantic encoding feature vector of the i-th municipal engineering supervision material text summary segment description and the self-supervised cluster center representation vector, (·) T represents the transpose of the eigenvector, s i (x i , v c ) represents the anti-correlation coefficient between the semantic encoding feature vector of the i-th municipal engineering supervision material text summary segment description and the self-supervised cluster center representation vector, sigmoid is the Sigmoid activation function, s a1 、s a2 、s ai and an represent the first, second, i-th and n-th normalized anti-correlation weight coefficients respectively, Xf A sequence representing the granular semantic feature vector of the anti-reinforcement market engineering supervision data segment, Indicates subtraction by position, X s Represents a sequence of segment-granular semantic feature vectors of the reverse attention selection market engineering supervision data.

[0077] Then, after the sequence of the reverse attention-selected market engineering supervision data segment granularity semantic feature vectors is cascaded into the market engineering supervision data text summary semantic salient coding vector, the market engineering supervision data text summary semantic salient coding vector is input into the classifier-based data classifier to obtain the classification result, and the classification result is the data type label of the municipal engineering supervision data to be managed. That is, first, the sequence of the reverse attention-selected market engineering supervision data segment granularity semantic feature vectors is feature integrated through the cascade operation to comprehensively represent the global semantic information of the municipal engineering supervision data and generate the market engineering supervision data text summary semantic salient coding vector. Then, it is input into the pre-trained classifier, and multi-classification prediction is performed according to the pre-set municipal engineering supervision data type label set to generate the probability distribution of the municipal engineering supervision data to be managed belonging to each type label, thereby determining the specific type of the data.

[0078] Correspondingly, in step S550, based on the sequence of the reverse attention-selected market engineering supervision data segment granularity semantic feature vectors, the classification result of the data type is determined, including: after cascading the sequence of the reverse attention-selected market engineering supervision data segment granularity semantic feature vectors into a market engineering supervision data text summary semantic salient coding vector, the market engineering supervision data text summary semantic salient coding vector is input into a classifier-based data classifier to obtain the classification result, and the classification result is the data type label of the municipal engineering supervision data to be managed.

[0079] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression and SVM are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification. However, this is prone to errors and is not efficient. Commonly used multi-classification methods include the Softmax classification function.

[0080] Furthermore, based on the classification result, the municipal engineering supervision data to be managed are archived.

[0081] In a preferred example, the market engineering supervision data segment significant information representation vector represents the segment-granular text significant features of the sequence of semantic coding feature vectors describing the municipal engineering supervision data text summary segment. Therefore, when the market engineering supervision data segment significant information representation vector is used as the self-supervised cluster center representation vector to perform feature reverse attention selection based on the center of the self-supervised cluster on the sequence of semantic coding feature vectors describing the municipal engineering supervision data text summary segment, the sequence of the reverse attention selected market engineering supervision data segment granularity semantic feature vectors will cause feature distribution semantic regression offset due to the reverse attention weight difference, which will lead to the market engineering supervision data text summary semantic significant coding vector obtained by cascading reverse attention selection of the sequence of market engineering supervision data segment granularity semantic feature vectors having insufficient offset semantic coverage relative to the source domain content text, thereby causing outlier class regression inference mapping deviation, affecting the accuracy of the classification result obtained by inputting the market engineering supervision data text summary semantic significant coding vector into the data classifier based on the classifier.

[0082] Based on this, when the semantically significant coding vector of the market engineering supervision document text summary is input into a classifier-based document classifier to obtain a classification result, the semantically significant coding vector of the market engineering supervision document text summary is optimized.

[0083] Specifically, the semantically significant coding vector of the market engineering supervision document text summary is input into a classifier-based document classifier to obtain the classification result, including: optimizing the semantically significant coding vector of the market engineering supervision document text summary to obtain an optimized semantically significant coding vector of the market engineering supervision document text summary; inputting the optimized semantically significant coding vector of the market engineering supervision document text summary into the classifier-based document classifier to obtain the classification result.

[0084] Among them, the semantically significant coding vector of the market engineering supervision document text summary is optimized to obtain the optimized semantically significant coding vector of the market engineering supervision document text summary, which specifically includes: calculating the feature mean of the semantically significant coding vector of the market engineering supervision document text summary, and dividing the feature mean by the difference between the maximum eigenvalue and the minimum eigenvalue of the semantically significant coding vector of the market engineering supervision document text summary to obtain the semantically significant coding distribution representation value of the market engineering supervision document text summary; subtracting the semantically significant coding distribution representation value of the market engineering supervision document text summary and dividing it by the semantically significant coding distribution representation value of the market engineering supervision document text summary to obtain the semantically significant coding distribution modulation value of the market engineering supervision document text summary; activating the semantically significant coding vector of the market engineering supervision document text summary through a probabilistic function to obtain a probabilistic semantically significant coding vector of the market engineering supervision document text summary; comparing the probabilistic semantically significant coding vector of the market engineering supervision document text summary with the market engineering supervision document text summary. After point-wise subtraction of the modulation value of the semantically significant coding distribution of the text summary of the supervision material, take the absolute value and calculate the negative of the logarithmic value with base 2 to obtain the probabilistic market engineering supervision material text summary semantically significant coding distribution modulation information vector; after dividing the semantically significant coding distribution representation value of the text summary of the market engineering supervision material by one minus the difference of each eigenvalue of the probabilistic market engineering supervision material text summary semantically significant coding vector, sum up all eigenvalues ​​of the probabilistic market engineering supervision material text summary semantically significant coding vector, and divide it by the length of the market engineering supervision material text summary semantically significant coding vector to obtain the probabilistic market engineering supervision material text summary semantically significant coding distribution modulation bias value; point-wise addition of the probabilistic market engineering supervision material text summary semantically significant coding distribution modulation information vector and the probabilistic market engineering supervision material text summary semantically significant coding distribution modulation bias value and the weight as a hyperparameter to obtain the optimized market engineering supervision material text summary semantically significant coding vector.

[0085] The optimization process of the semantically significant encoding vector of the market engineering supervision material text summary is expressed as follows:

[0086]

[0087] And among them:

[0088]

[0089] in, represents the feature mean of the semantically significant encoding vector of the market engineering supervision document text summary, v max and v minrepresents the maximum eigenvalue and the minimum eigenvalue in the semantically significant coding vector of the market engineering supervision document text summary, respectively; p represents the semantically significant coding distribution representation value of the market engineering supervision document text summary; V represents the probabilistic semantically significant coding vector of the market engineering supervision document text summary obtained after the semantically significant coding vector of the market engineering supervision document text summary is activated by a probabilistic function; v i represents the ith eigenvalue of the semantically significant encoding vector of the probabilistic market engineering supervision document text summary, log represents the logarithmic function value with base 2, ε is the weight as a hyperparameter, the value of L is the length of the semantically significant encoding vector of the probabilistic market engineering supervision document text summary, V ′ Represents the semantically significant encoding vector of the optimized market engineering supervision material text summary.

[0090] That is, in the above preferred example, the eigenvalue-based probability information distribution planning of the semantically significant coding vector of the text summary of the market engineering supervision materials is carried out through the Bernoulli probability modulation distribution of the semantically significant coding vector of the text summary of the market engineering supervision materials relative to the eigenvalue distribution, and the probability reverse mapping of the probability characteristics of the semantically significant coding vector of the text summary of the market engineering supervision materials is used as the extended coverage of the set mapping space of the semantically significant coding vector of the text summary of the market engineering supervision materials, so as to independently understand the interaction path between the intuitive probability information distribution and the abstract probability space mapping of the semantically significant coding vector of the text summary of the market engineering supervision materials, so as to improve the accuracy of the classification result obtained by the data classifier based on the classifier of the optimized semantically significant coding vector of the text summary of the market engineering supervision materials by avoiding the counterfactual reasoning mapping of the outlier feature distribution of the semantically significant coding vector of the text summary of the market engineering supervision materials to the class regression probability.

[0091] Based on the above embodiments, see Figure 5As shown, it is a structural schematic diagram of a standardized management system 800 for municipal engineering supervision data in an embodiment of the present application. The standardized management system 800 for municipal engineering supervision data includes: a data collection module 810, which is used to obtain the municipal engineering supervision data to be managed; a summary generation module 820, which is used to generate a summary of the municipal engineering supervision data to be managed to obtain a text summary of the municipal engineering supervision data; a semantic encoding module 830, which is used to perform segment-based semantic encoding on the text summary of the municipal engineering supervision data to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description; a feature reverse attention selection module 840, which is used to perform feature reverse attention selection on the sequence of semantic encoding feature vectors of the municipal engineering supervision data text summary segment description to obtain a sequence of reverse attention selection market engineering supervision data segment granularity semantic feature vectors; a data type classification module 850, which is used to determine the classification result of the data type based on the sequence of reverse attention selection market engineering supervision data segment granularity semantic feature vectors; and an archiving module 860, which is used to archive the municipal engineering supervision data to be managed based on the classification result.

[0092] Here, those skilled in the art can understand that the specific functions and operations of each module in the above-mentioned municipal engineering supervision data standardization management system 800 have been described in the above reference. Figures 2 to 4 It has been introduced in detail in the description of the standardized management method of municipal engineering supervision data, and therefore, its repeated description will be omitted.

[0093] Figure 6 FIG. 1 is an application scenario diagram of the standardized management method for municipal engineering supervision data according to an embodiment of the present application. Figure 6 As shown, in this application scenario, first, obtain the supervision data of the municipal engineering project to be managed (for example, Figure 6 D), then, the municipal engineering supervision data to be managed is input into a server (for example, Figure 6 In S) shown in , the server is able to use the municipal engineering supervision data standardization management algorithm to process the municipal engineering supervision data to be managed to obtain a classification result of the data type label used to represent the municipal engineering supervision data to be managed.

[0094] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the standardized management method of municipal engineering supervision data in the above embodiments when executing the program.

[0095] For example, in the case of electronic equipment Figure 1 Taking the server 100 in the example as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.

[0096] The embodiment of the present application also provides a computer-readable storage medium, on which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, the standardized management method for municipal engineering supervision data according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0097] The embodiment of the present application also provides a computer program product or computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the municipal engineering supervision data standardization management method according to the embodiment of the present application.

[0098] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0099] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are only illustrative, and different aspects of the system and method can use different units.

[0100] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.

[0101] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0102] The above is an explanation of the present application and should not be considered as limiting thereof. Although several exemplary embodiments of the present application have been described, those skilled in the art will readily appreciate that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application.

Claims

1. A method for standardizing the management of municipal engineering supervision data, characterized in that: include: Obtain supervision information of municipal engineering projects to be managed; Generate a summary of the municipal engineering supervision data to be managed to obtain a text summary of the municipal engineering supervision data; Performing segment-granularity-based semantic coding on the municipal engineering supervision document text summary to obtain a sequence of segment description semantic coding feature vectors of the municipal engineering supervision document text summary; Performing feature reverse attention selection on the sequence of semantic coding feature vectors describing the text summary segment of the municipal engineering supervision data to obtain a sequence of reverse attention selection market engineering supervision data segment granularity semantic feature vectors; Determine the classification result of the data type based on the sequence of the granularity semantic feature vectors of the reverse attention selection market engineering supervision data segment; Based on the classification result, the municipal engineering supervision data to be managed are archived.

2. The method for standardizing the management of municipal engineering supervision data according to claim 1 is characterized in that: The summary of the municipal engineering supervision data to be managed is generated to obtain a text summary of the municipal engineering supervision data, including: The municipal engineering supervision data to be managed is input into a text summary generator based on the AIGC model to obtain a text summary of the municipal engineering supervision data.

3. The method for standardizing the management of municipal engineering supervision data according to claim 2 is characterized in that: The municipal engineering supervision document text summary is semantically encoded based on segment granularity to obtain a sequence of municipal engineering supervision document text summary segment description semantic encoding feature vectors, including: Segment processing is performed on the municipal engineering supervision document text summary to obtain a sequence of municipal engineering supervision document text summary segment descriptions; Semantic coding is performed on each municipal engineering supervision document text summary segment description in the sequence of the municipal engineering supervision document text summary segment description to obtain a sequence of semantic coding feature vectors of the municipal engineering supervision document text summary segment description.

4. The method for standardizing the management of municipal engineering supervision data according to claim 3 is characterized in that: Semantically encoding each municipal engineering supervision document text summary segment description in the sequence of municipal engineering supervision document text summary segment descriptions to obtain a sequence of semantic encoding feature vectors of the municipal engineering supervision document text summary segment descriptions, including: Each municipal engineering supervision document text summary segment description in the sequence of the municipal engineering supervision document text summary segment description is input into the supervision document semantic understander based on the Bert model to obtain a sequence of the municipal engineering supervision document text summary segment description semantic encoding feature vectors.

5. The method for standardizing the management of municipal engineering supervision data according to claim 4 is characterized in that: The sequence of semantic coding feature vectors describing the text summary segment of the municipal engineering supervision data is subjected to feature reverse attention selection to obtain a sequence of reverse attention selected market engineering supervision data segment granularity semantic feature vectors, including: Extracting significant features from the sequence of semantically encoded feature vectors describing the text summary segment of the municipal engineering supervision data to obtain a significant information representation vector of the municipal engineering supervision data segment; Taking the significant information representation vector of the market engineering supervision data segment as the self-supervised cluster center representation vector, the sequence of semantic encoding feature vectors describing the text summary segment of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain a sequence of reverse attention selected granular semantic feature vectors of the market engineering supervision data segment.

6. The method for standardizing the management of municipal engineering supervision data according to claim 5 is characterized in that: Extracting significant features from the sequence of semantic coding feature vectors describing the text summary segment of the municipal engineering supervision data to obtain a significant information representation vector of the municipal engineering supervision data segment includes: The maximum values ​​of the semantic coding feature vectors of each municipal engineering supervision data text summary segment description in the sequence of the semantic coding feature vectors of the municipal engineering supervision data text summary segment description are extracted to obtain the municipal engineering supervision data segment significant information representation vector composed of multiple maximum values.

7. The method for standardizing the management of municipal engineering supervision data according to claim 6 is characterized in that: Taking the significant information representation vector of the market engineering supervision data segment as the self-supervised cluster center representation vector, the sequence of the semantic encoding feature vectors of the text summary segment description of the municipal engineering supervision data is subjected to feature reverse attention selection based on the self-supervised cluster center to obtain the sequence of the reverse attention selected market engineering supervision data segment granular semantic feature vectors, including: Calculating the anti-correlation coefficient between each municipal engineering supervision document text summary segment description semantic coding feature vector in the sequence of the municipal engineering supervision document text summary segment description semantic coding feature vector and the self-supervised cluster center representation vector to obtain a sequence of anti-correlation coefficients; Inputting the sequence of anti-correlation coefficients into a Sigmoid activation function to obtain a sequence of normalized anti-correlation weight coefficients; Using each normalized anti-correlation weight coefficient in the sequence of normalized anti-correlation weight coefficients as a weight, weighting each municipal engineering supervision material text summary segment description semantic coding feature vector in the sequence of municipal engineering supervision material text summary segment description semantic coding feature vectors respectively to obtain a sequence of anti-reinforcement market engineering supervision material segment granularity semantic feature vectors; Calculate the positional difference between each group of corresponding municipal engineering supervision data text summary segment description semantic coding feature vectors and anti-reinforcement market engineering supervision data segment granularity semantic feature vectors in the sequence of the municipal engineering supervision data text summary segment description semantic coding feature vectors and the sequence of the anti-reinforcement market engineering supervision data segment granularity semantic feature vectors to obtain the sequence of the reverse attention selection market engineering supervision data segment granularity semantic feature vectors.

8. The method for standardizing the management of municipal engineering supervision data according to claim 7 is characterized in that: Calculating the anti-correlation coefficient between each municipal engineering supervision document text summary segment description semantic coding feature vector in the sequence of the municipal engineering supervision document text summary segment description semantic coding feature vector and the self-supervised cluster center representation vector to obtain a sequence of anti-correlation coefficients, including: Calculate the covariance matrix between the semantic encoding feature vector of the municipal engineering supervision material text summary segment description and the self-supervised cluster center representation vector; Calculate the position-by-position subtraction between the semantic encoding feature vector of the municipal engineering supervision material text summary segment description and the self-supervised cluster center representation vector to obtain a semantic differential feature vector; The anti-correlation coefficient is obtained by calculating the square root of a product result of a transposed vector of the semantic differential feature vector, the covariance matrix, and the semantic differential feature vector.

9. The method for standardizing the management of municipal engineering supervision data according to claim 8 is characterized in that: Based on the reverse attention selection market engineering supervision data segment granularity semantic feature vector sequence, determine the data type classification result, including: After the sequence of reverse attention-selected market engineering supervision data segment-granularity semantic feature vectors is cascaded into a market engineering supervision data text summary semantic salient coding vector, the market engineering supervision data text summary semantic salient coding vector is input into a classifier-based data classifier to obtain the classification result, which is the data type label of the municipal engineering supervision data to be managed.

Images