[go: up one dir, main page]

CN118154135B - Flow configuration system and method applied to development operation and maintenance management platform - Google Patents

Flow configuration system and method applied to development operation and maintenance management platform Download PDF

Info

Publication number
CN118154135B
CN118154135B CN202410565476.3A CN202410565476A CN118154135B CN 118154135 B CN118154135 B CN 118154135B CN 202410565476 A CN202410565476 A CN 202410565476A CN 118154135 B CN118154135 B CN 118154135B
Authority
CN
China
Prior art keywords
semantic
sentence
training
sequence
coding
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410565476.3A
Other languages
Chinese (zh)
Other versions
CN118154135A (en
Inventor
倪杨喜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hangzhou Dingdang Technology Co ltd
Original Assignee
Hangzhou Dingdang Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hangzhou Dingdang Technology Co ltd filed Critical Hangzhou Dingdang Technology Co ltd
Priority to CN202410565476.3A priority Critical patent/CN118154135B/en
Publication of CN118154135A publication Critical patent/CN118154135A/en
Application granted granted Critical
Publication of CN118154135B publication Critical patent/CN118154135B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G06Q10/103—Workflow collaboration or project management
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/10—Text processing
    • G06F40/12—Use of codes for handling textual entities
    • G06F40/126—Character encoding
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F8/00—Arrangements for software engineering
    • G06F8/10—Requirements analysis; Specification techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F8/00—Arrangements for software engineering
    • G06F8/20—Software design
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/042—Knowledge-based neural networks; Logical representations of neural networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Computing Systems (AREA)
  • Business, Economics & Management (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Machine Translation (AREA)

Abstract

The application relates to the technical field of intelligent recommendation, in particular to a flow configuration system and a method applied to an development operation and maintenance management platform, which utilize an artificial intelligence technology based on deep learning to carry out semantic analysis based on sentence granularity on business demand information input by a user, the sentence granularity semantic features of the business requirement information are captured, and the sentence semantic association coding is carried out on the business requirement information based on the sentence semantic adjacency relationship of the business requirement information so as to fully mine the global semantic features of the business requirement information, thereby intelligently recommending a proper flow basic structure according to the business requirement of a user. Therefore, the operation and maintenance efficiency is effectively improved, the operation and maintenance cost is reduced, and a more intelligent and efficient development operation and maintenance management platform is provided for enterprises.

Description

Flow configuration system and method applied to development operation and maintenance management platform
Technical Field
The application relates to the technical field of intelligent recommendation, in particular to a flow configuration system and method applied to an development operation and maintenance management platform.
Background
With the rapid development of information technology, IT architecture of enterprises is increasingly complex, and requirements on operation and maintenance management are also higher. In this context, development of operation and maintenance management platforms (DevOps) has been developed. The Development operation and maintenance management platform is a comprehensive platform integrating Development and operation functions, and aims to break the barriers between Development and operation and maintenance and realize seamless butt joint and efficient cooperation between the Development and operation. By developing the operation and maintenance management platform, a development team can rapidly deploy applications, monitor the running condition of the system and timely respond to various faults, so that the speed and quality of software delivery are improved.
The process configuration is a core component of the development operation and maintenance management platform, and refers to the process of planning, designing, implementing and managing the operation and maintenance process according to service requirements, organization structures and resource conditions in the development operation and maintenance management platform. The main goal of the flow configuration is to realize efficient and standardized operation and maintenance operation, and ensure stable operation and continuous delivery of the system. However, the conventional process configuration manner often requires a user to have a high technical level to perform configuration effectively, which brings a certain challenge to the operation and maintenance management of the enterprise.
Therefore, an optimized flow configuration system and method for developing an operation and maintenance management platform are desired.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides a flow configuration system and a flow configuration method applied to an development operation and maintenance management platform, which utilize an artificial intelligence technology based on deep learning to carry out semantic analysis based on sentence granularity on business requirement information input by a user, capture sentence granularity semantic features of the business requirement information, and carry out inter-sentence semantic association coding on the business requirement information based on inter-sentence semantic adjacency relations of the business requirement information so as to fully extract global semantic features of the business requirement information, thereby intelligently recommending a proper flow basic structure according to the business requirement of the user. Therefore, the operation and maintenance efficiency is effectively improved, the operation and maintenance cost is reduced, and a more intelligent and efficient development operation and maintenance management platform is provided for enterprises.
Accordingly, according to one aspect of the present application, there is provided a flow configuration system applied to an development operation and maintenance management platform, comprising:
the business requirement acquisition module is used for acquiring text description of business requirements input by a user;
the sentence-granularity semantic coding module is used for carrying out sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of the sentence-granularity semantic coding feature vectors of the service requirement description;
the inter-sentence semantic association coding module is used for carrying out inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic association coding feature vectors so as to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors;
The feature screening module is used for extracting essential features of the sequence of the semantic association coding feature vectors among the business requirement description sentences to obtain business requirement semantic understanding feature vectors;
And the flow basic structure recommending module is used for determining the recommending type of the flow basic structure based on the service demand semantic understanding feature vector.
In the above process configuration system applied to the development operation and maintenance management platform, the sentence granularity semantic coding module includes: the clause processing unit is used for carrying out clause processing on the text description of the service requirement to obtain a sequence of the service requirement description sentence; the semantic coding unit is used for respectively carrying out semantic coding on each business requirement description sentence in the sequence of the business requirement description sentences by using a semantic coder based on a transducer model so as to obtain the sequence of the business requirement description sentence granularity semantic coding feature vector.
In the above process configuration system applied to the development operation and maintenance management platform, the inter-sentence semantic association coding module includes: the semantic adjacency topology matrix construction unit is used for calculating the semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors; the semantic adjacency topological feature extraction unit is used for enabling the semantic adjacency topological matrix to pass through a semantic adjacency topological feature extractor based on a convolutional neural network model so as to obtain a semantic adjacency topological feature matrix; the semantic association coding unit is used for enabling the sequence of the business requirement description sentence granularity semantic association coding feature vectors and the semantic adjacency topological feature matrix to pass through an inter-sentence semantic association coder based on a graph convolution neural network model so as to obtain the sequence of the business requirement description inter-sentence semantic association coding feature vectors.
In the above flow configuration system applied to the development operation and maintenance management platform, the semantic adjacency topology matrix construction unit is configured to: calculating a semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors by using the following semantic adjacency association formula; the semantic adjacency association formula is as follows:
Wherein, Is the first in the sequence of the business requirement description sentence granularity semantic coding feature vectorsThe individual business requirement descriptors granularity semantic coding feature vectors,Is the j-th business requirement description sentence granularity semantic coding feature vector,Is the firstStandard deviation between the granularity semantic coding feature vector of the individual business requirement description sentence and the granularity semantic coding feature vector of the jth business requirement description sentence,Representing the square of the 2 norms of the feature vectors,The element pair bit subtracting difference processing is represented,Representing the operation of a natural exponential function,For the first of the semantic adjacency topology matricesThe characteristic value of the location is used to determine,And the number of feature vectors is encoded for the granularity semantics of the business requirement description sentence.
In the above process configuration system applied to the development operation and maintenance management platform, the feature screening module is configured to: and the sequence of the semantic association coding feature vectors among the business requirement description sentences is subjected to semantic gating-based essential feature attention selection network to obtain the business requirement semantic understanding feature vectors.
In the above process configuration system applied to the development operation and maintenance management platform, the feature screening module is configured to: processing the sequence of the semantic association coding feature vectors among the business requirement description sentences by using the following essential feature attention selection formula to obtain the business requirement semantic understanding feature vectors; wherein, the essential feature attention selection formula is:
Wherein, Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Representing the 1-norm of the feature vector,Describing the length-1 of the sequence of inter-sentence semantic association encoded feature vectors for the business requirement,A representation of a sequence of inter-sentence semantic association encoded feature vectors is described for the business requirement,The attention is given a score by a factor,The masking process is represented by a process of masking,Scoring coefficients for masking the attention,Representing the operation of a natural exponential function,Representing the total number of masked attention scoring coefficients,And semantic understanding of feature vectors for the business requirements.
In the above process configuration system applied to the development operation and maintenance management platform, the process basic structure recommendation module is configured to: and passing the service demand semantic understanding feature vector through a flow basic structure recommender based on a classifier to obtain a classification result, wherein the classification result is used for representing type labels of the flow basic structure. The flow configuration system applied to the development operation and maintenance management platform further comprises a training module for training the semantic encoder based on the transducer model, the semantic adjacency topological feature extractor based on the convolutional neural network model, the inter-sentence semantic association encoder based on the graph convolutional neural network model, the essential feature attention selection network based on semantic gating and the flow basic structure recommender based on the classifier.
In the above process configuration system applied to the development operation and maintenance management platform, the training module includes: the training data acquisition unit is used for acquiring training data, wherein the training data comprises text description of training service requirements input by a user and a true value of a flow basic structure type; the training data clause unit is used for carrying out clause processing on the text description of the training service requirement to obtain a sequence of training service requirement description sentences; the training data semantic coding unit is used for respectively carrying out semantic coding on each training service demand description sentence in the sequence of the training service demand description sentences by using the semantic coder based on the Transformer model so as to obtain a sequence of training service demand description sentence granularity semantic coding feature vectors; the training data semantic adjacency correlation unit is used for calculating a training semantic adjacency topology matrix of the sequence of the training business requirement description sentence granularity semantic coding feature vectors; the training data semantic adjacency feature extraction unit is used for enabling the training semantic adjacency topological matrix to pass through the semantic adjacency topological feature extractor based on the convolutional neural network model so as to obtain a training semantic adjacency topological feature matrix; the training data semantic association unit is used for enabling the training service requirement description sentence granularity semantic coding feature vector sequence and the training semantic adjacency topological feature matrix to pass through the inter-sentence semantic association encoder based on the graph convolution neural network model so as to obtain the training service requirement description inter-sentence semantic association coding feature vector sequence; the training data feature selection unit is used for enabling the sequence of the training service requirement description inter-sentence semantic association coding feature vectors to pass through the essential feature attention selection network based on semantic gating to obtain training service requirement semantic understanding feature vectors; the classification loss unit is used for passing the training service demand semantic understanding feature vector through the classifier-based flow basic structure recommender to obtain a classification loss function value; the model training unit is used for training the semantic encoder based on the transform model, the semantic adjacency topological feature extractor based on the convolutional neural network model, the inter-sentence semantic association encoder based on the graph convolutional neural network model, the semantic gating-based essential feature attention selection network and the classifier-based flow basic structure recommender by the classification loss function value, wherein in each iteration of training, the training service requirement semantic understanding feature vector is optimized. According to another aspect of the present application, there is provided a flow configuration method applied to an development operation and maintenance management platform, including:
acquiring text description of business requirements input by a user;
Performing sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of sentence-granularity semantic coding feature vectors of the service requirement description;
Performing inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic association coding feature vectors to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors;
Extracting essential characteristics of the sequence of the semantic association coding feature vectors among the business requirement description sentences to obtain business requirement semantic understanding feature vectors; and determining the recommended type of the basic flow structure based on the semantic understanding feature vector of the business requirement.
Compared with the prior art, the flow configuration system and method applied to the development operation and maintenance management platform provided by the application have the advantages that the artificial intelligence technology based on deep learning is utilized to carry out semantic analysis based on sentence granularity on business demand information input by a user, the sentence granularity semantic features of the business demand information are captured, and the sentence semantic association coding is carried out on the business demand information based on the sentence semantic adjacency relationship of the business demand information, so that the global semantic features of the business demand information are fully extracted, and therefore, a proper flow basic structure is intelligently recommended according to the business demand of the user. Therefore, the operation and maintenance efficiency is effectively improved, the operation and maintenance cost is reduced, and a more intelligent and efficient development operation and maintenance management platform is provided for enterprises.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing embodiments of the present application in more detail with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and together with the embodiments of the application, and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
Fig. 1 is a block diagram of a flow configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application.
Fig. 2 is a schematic architecture diagram of a flow configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application.
Fig. 3 is a block diagram of a sentence granularity semantic coding module in a process configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application.
Fig. 4 is a block diagram of an inter-sentence semantic association encoding module in a process configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application.
Fig. 5 is a block diagram of a training module in a process configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application.
Fig. 6 is a flowchart of a flow configuration method applied to an development operation and maintenance management platform according to an embodiment of the present application.
Detailed Description
The foregoing and other objects, features and advantages of the application will be apparent from the following more particular description of embodiments of the application, as illustrated in the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein. Meanwhile, the accompanying drawings are included to provide a further understanding of embodiments of the application, and constitute a part of this specification, illustrate the application and together with the embodiments of the application, and not to limit the application. In the drawings, like reference numerals generally refer to like parts or steps. Fig. 1 is a block diagram of a flow configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application. Fig. 2 is a schematic architecture diagram of a flow configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application. As shown in fig. 1 and 2, a process configuration system 100 applied to an development operation and maintenance management platform according to an embodiment of the present application includes: a service requirement acquisition module 110, configured to acquire a text description of a service requirement input by a user; the sentence-granularity semantic coding module 120 is configured to perform sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of sentence-granularity semantic coding feature vectors of the service requirement description; the inter-sentence semantic association coding module 130 is configured to perform inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic coding feature vectors to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors; the feature screening module 140 is configured to extract essential features of the sequence of the semantic association coding feature vectors between the service requirement description sentences to obtain service requirement semantic understanding feature vectors; the process basic structure recommending module 150 is configured to determine a recommended type of the process basic structure based on the service requirement semantic understanding feature vector. As described above in the background art, the conventional flow configuration manner often requires a user to have a higher technical level to perform configuration effectively, which brings a certain challenge to the operation and maintenance management of the enterprise. Aiming at the technical problems, the technical concept of the application is to utilize an artificial intelligence technology based on deep learning to carry out semantic analysis based on sentence granularity on business requirement information input by a user, capture sentence granularity semantic features of the business requirement information, and carry out inter-sentence semantic association coding on the business requirement information based on inter-sentence semantic adjacency relations of the business requirement information so as to fully extract global semantic features of the business requirement information, thereby intelligently recommending a proper flow basic structure according to the business requirement of the user. Therefore, the operation and maintenance efficiency is effectively improved, the operation and maintenance cost is reduced, and a more intelligent and efficient development operation and maintenance management platform is provided for enterprises.
In the above-mentioned process configuration system 100 applied to the development operation and maintenance management platform, the service requirement obtaining module 110 is configured to obtain a text description of a service requirement input by a user. It should be appreciated that the textual description of the business needs entered by the user is a significant source of data that directly reflects their intent and expectations. In the technical scheme of the application, the text description of the service requirement is subjected to semantic analysis and processing, so that the requirement and intention of a user can be accurately understood, personalized customization is performed according to the specific service requirement of the user, and a proper flow basic structure is recommended for the user, so that the user can more conveniently create, edit and manage each link of the flow, and the requirements of different service scenes are met.
In the above-mentioned process configuration system 100 applied to the development operation and maintenance management platform, the sentence-granularity semantic coding module 120 is configured to perform sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of sentence-granularity semantic coding feature vectors of the service requirement description. Specifically, fig. 3 is a block diagram of a sentence granularity semantic encoding module in a process configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application. As shown in fig. 3, the sentence-granularity semantic encoding module 120 includes: a clause processing unit 121, configured to perform clause processing on the text description of the service requirement to obtain a sequence of service requirement description sentences; The semantic coding unit 122 is configured to perform semantic coding on each service requirement description sentence in the sequence of service requirement description sentences by using a semantic coder based on a transform model, so as to obtain a sequence of granularity semantic coding feature vectors of the service requirement description sentences. Specifically, the clause processing unit 121 is configured to process the text description of the service requirement to obtain a sequence of the service requirement description sentence. It should be appreciated that, considering that the text description of the business requirement is usually a long text, the long text usually has complex structure and logic relationship, and the context in the long text may change with the increase of the text length, if the whole text is directly encoded, the local context information may be lost, and the structural information inside the text cannot be sufficiently captured, which affects the accurate understanding of the whole text semantics. Therefore, in order to perform finer semantic understanding on the business requirements of the users, sentence segmentation is further performed on the text description of the business requirements so as to decompose the text description of the business requirements into separate sentences from the text description of the long sentences, each sentence is processed separately more efficiently, semantic features of semantic units of each sentence are captured more finely, the user requirements are understood more accurately, and efficiency and performance of data semantic coding are improved. Specifically, the semantic coding unit 122 is configured to use a semantic encoder based on a transform model to perform semantic coding on each service requirement description sentence in the sequence of service requirement description sentences to obtain a sequence of granularity semantic coding feature vectors of the service requirement description sentences. Those of ordinary skill in the art will appreciate that a transducer is a natural language processing model that captures long-range dependencies between words in text data based entirely on the attentional mechanisms, discarding traditional neural network structures such as Recurrent Neural Networks (RNNs) and Convolutional Neural Networks (CNNs). In the technical scheme of the application, the semantic encoder based on the transducer model carries out parallel calculation on each word in the business requirement description sentence, and captures semantic dependency relations among each word in the business requirement description sentence by using a self-attention mechanism so as to realize context semantic attention in the sentence, thereby being capable of better understanding the semantic structure of each business requirement description sentence. In addition, semantic coding is respectively carried out on each business requirement description sentence in the mode, so that the fine semantic representation can be realized, the sentence independence is maintained, the semantic interference among different sentences is avoided, and the accuracy of semantic coding is improved.
In the above-mentioned process configuration system 100 applied to the development operation and maintenance management platform, the inter-sentence semantic association coding module 130 is configured to perform inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic coding feature vectors to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors. Specifically, fig. 4 is a block diagram of an inter-sentence semantic association encoding module in a process configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application. As shown in fig. 4, the inter-sentence semantic association encoding module 130 includes: a semantic adjacency topology matrix construction unit 131, configured to calculate a semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors; a semantic adjacent topological feature extraction unit 132, configured to pass the semantic adjacent topological matrix through a semantic adjacent topological feature extractor based on a convolutional neural network model to obtain a semantic adjacent topological feature matrix; the semantic association encoding unit 133 is configured to obtain the sequence of the semantic association encoding feature vectors between the business requirement description sentences by using an inter-sentence semantic association encoder based on a graph convolution neural network model with the sequence of the semantic association encoding feature vectors between the business requirement description sentences and the semantic adjacency topological feature matrix. Specifically, the semantic adjacency topology matrix construction unit 131 is configured to calculate a semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors. It should be appreciated that, considering that there is some semantic association between the business requirement descriptions, it is desirable to capture the semantic association between the business requirement descriptions so as to more fully understand the overall semantic structure of the text description of the business requirement. Based on the above, in the technical scheme of the application, the semantic association degree among the business requirement description sentences is quantified by calculating the semantic adjacency topological matrix of the sequence of the business requirement description sentence granularity semantic coding feature vector, and the context relation among the business requirement description sentences is modeled, so that the context and the association of the business requirement description sentences in the whole text description are better understood, and the semantic background of the user business requirement description is more accurately grasped. In a specific example of the present application, the semantic adjacency topology matrix construction unit 131 is configured to: calculating a semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors by using the following semantic adjacency association formula; the semantic adjacency association formula is as follows:
Wherein, Is the first in the sequence of the business requirement description sentence granularity semantic coding feature vectorsThe individual business requirement descriptors granularity semantic coding feature vectors,Is the j-th business requirement description sentence granularity semantic coding feature vector,Is the firstStandard deviation between the granularity semantic coding feature vector of the individual business requirement description sentence and the granularity semantic coding feature vector of the jth business requirement description sentence,Representing the square of the 2 norms of the feature vectors,The element pair bit subtracting difference processing is represented,Representing the operation of a natural exponential function,For the first of the semantic adjacency topology matricesThe characteristic value of the location is used to determine,And the number of feature vectors is encoded for the granularity semantics of the business requirement description sentence. Specifically, the semantic adjacent topological feature extraction unit 132 is configured to pass the semantic adjacent topological matrix through a semantic adjacent topological feature extractor based on a convolutional neural network model to obtain a semantic adjacent topological feature matrix. It should be understood that, in order to learn and extract meaningful inter-sentence semantic association features from the semantic adjacency topology matrix, in the technical solution of the present application, a semantic adjacency topology feature extractor based on a convolutional neural network model is used to process the semantic adjacency topology matrix. It will be appreciated by those of ordinary skill in the art that Convolutional Neural Network (CNN) is a feed-forward neural network with deep architecture, with strong feature extraction capabilities. according to the technical scheme, the semantic adjacency topological feature extractor based on the convolutional neural network model can effectively learn and extract local semantic association features among sentences in the adjacency topological matrix by performing sliding convolution calculation on the semantic adjacency topological matrix, so that semantic connection relations among all business requirement description sentences can be better understood. In addition, the CNN model can better fit the complex semantic relation in the semantic adjacency topology matrix by introducing a nonlinear activation function, so that the semantic understanding capability of the user service requirement is improved. Specifically, the semantic association encoding unit 133 is configured to pass the sequence of the business requirement description sentence granularity semantic association encoding feature vectors and the semantic adjacency topological feature matrix through an inter-sentence semantic association encoder based on a graph convolution neural network model to obtain the sequence of the business requirement description inter-sentence semantic association encoding feature vectors. That is, semantic association coding is performed on semantic features of each business requirement description sentence based on semantic connection relations among each business requirement description sentence, so that semantic information is transferred among each business requirement description sentence, and the overall understanding capability of the business requirement information of the user is improved. Specifically, in the technical scheme of the application, an inter-sentence semantic association encoder based on a graph convolutional neural network model is used for carrying out feature transfer fusion encoding on the sequence of the semantic coding feature vectors of the service requirement description sentence granularity and the semantic adjacent topological feature matrix. It should be appreciated that the graph convolution neural network is a deep learning model specifically used to process graph structure data, and is capable of effectively capturing spatial dependencies in the graph structure and learning complex associated feature representations between nodes. In the technical scheme of the application, the sequence of the semantic coding feature vector with the granularity of the business requirement description sentence is used as each node in the graph structure data, and the semantic adjacency topological feature matrix is used for describing adjacency relations among each node, namely edges in the graph structure data. The feature transfer fusion coding is carried out on the two business requirement description sentences through the graph convolution neural network, semantic features of the business requirement description sentences and semantic association among the sentences can be considered at the same time, and feature updating of the business requirement description sentence granularity semantic coding feature vector is realized through aggregation and transfer of features of neighborhood nodes in a graph structure, so that semantic representation capacity of the business requirement description sentences is improved based on the inter-sentence semantic association features of text description of the business requirements.
In the above-mentioned process configuration system 100 applied to the development operation and maintenance management platform, the feature screening module 140 is configured to extract essential features of the sequence of the semantic association coding feature vectors between the business requirement description sentences to obtain the business requirement semantic understanding feature vectors. In a specific example of the present application, the method for extracting essential features of the sequence of semantic association coding feature vectors between the business requirement description sentences to obtain the business requirement semantic understanding feature vector is to obtain the business requirement semantic understanding feature vector by selecting a network based on semantic gating essential feature attention from the sequence of semantic association coding feature vectors between the business requirement description sentences. It should be appreciated that in order to obtain an overall semantic feature representation of the textual description of the business need, feature integration of the sequence of inter-sentence semantic association encoded feature vectors of the business need description is required. At the same time, the feature information in the sequence of inter-sentence semantic association encoded feature vectors of the business requirement description is not all useful in view of the overall semantic understanding of the textual description of the business requirement, which may contain some interference or redundant information. Therefore, in order to improve the overall semantic understanding capability of the business requirement text description, the information refining and screening are further carried out on the sequence of the semantic association coding feature vectors among the business requirement description sentences through a semantic gating-based essential feature attention selection network so as to remove redundant information in the business requirement description sentence, and important information is reserved. Specifically, the semantic association between features can be effectively modeled by the semantic gating-based essential feature attention selection network, weights of different features are dynamically adjusted in a feature sequence through an attention selection mechanism, and attention weighted fusion is carried out on the sequence of semantic association coding feature vectors among business requirement description sentences, so that the features with the most representation and criticality are automatically learned and selected, and the semantic representation and understanding capability of the business requirement description is improved.
Specifically, the feature screening module 140 is configured to: processing the sequence of the semantic association coding feature vectors among the business requirement description sentences by using the following essential feature attention selection formula to obtain the business requirement semantic understanding feature vectors; wherein, the essential feature attention selection formula is:
Wherein, Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Representing the 1-norm of the feature vector,Describing the length-1 of the sequence of inter-sentence semantic association encoded feature vectors for the business requirement,A representation of a sequence of inter-sentence semantic association encoded feature vectors is described for the business requirement,The attention is given a score by a factor,The masking process is represented by a process of masking,Scoring coefficients for masking the attention,Representing the operation of a natural exponential function,Representing the total number of masked attention scoring coefficients,And semantic understanding of feature vectors for the business requirements.
In the above-mentioned process configuration system 100 applied to the development operation and maintenance management platform, the process basic structure recommendation module 150 is configured to determine a recommendation type of the process basic structure based on the service requirement semantic understanding feature vector. In a specific example of the present application, the implementation manner of determining the recommendation type of the flow basic structure based on the service requirement semantic understanding feature vector is to pass the service requirement semantic understanding feature vector through a classifier-based flow basic structure recommender to obtain a classification result, where the classification result is used to represent a type tag of the flow basic structure. It should be appreciated that the recommendation of the flow infrastructure is essentially a classification problem. In the technical scheme of the application, the flow basic structure recommender performs feature learning and classification mapping on the service demand semantic understanding feature vector based on a classification algorithm so as to map the service demand semantic understanding feature vector to a specific type label of a flow basic structure, thereby recommending a proper flow basic structure for the service demand semantic understanding feature according to the semantic understanding feature of the user service demand, enabling the user to more conveniently create, edit and manage each link of the flow, and further improving the accuracy and efficiency of flow configuration.
It should be appreciated that prior to utilizing the above neural network model, training of the transfomer model-based semantic encoder, the convolutional neural network model-based semantic adjacency topological feature extractor, the graph convolutional neural network model-based inter-sentence semantic association encoder, the semantic gating-based essential feature attention selection network, and the classifier-based flow infrastructure recommender is required. That is, in the process configuration system applied to the development operation and maintenance management platform, the application further comprises a training module for training the semantic encoder based on the transform model, the semantic adjacency topological feature extractor based on the convolutional neural network model, the inter-sentence semantic association encoder based on the graph convolutional neural network model, the essential feature attention selection network based on semantic gating and the flow basic structure recommender based on the classifier.
Fig. 5 is a block diagram of a training module in a process configuration system applied to an development operation and maintenance management platform according to an embodiment of the present application. As shown in fig. 5, the training module 200 includes: a training data obtaining unit 210, configured to obtain training data, where the training data includes a text description of a training service requirement input by a user, and a true value of a process basic structure type; a training data clause unit 220, configured to process the text description of the training service requirement to obtain a sequence of training service requirement description sentences; the training data semantic coding unit 230 is configured to use the semantic encoder based on the transform model to perform semantic coding on each training service requirement description sentence in the sequence of training service requirement description sentences to obtain a sequence of training service requirement description sentence granularity semantic coding feature vectors; a training data semantic adjacency correlation unit 240, configured to calculate a training semantic adjacency topology matrix of the sequence of training service requirement description sentence granularity semantic coding feature vectors; a training data semantic adjacency feature extraction unit 250, configured to pass the training semantic adjacency topology matrix through the semantic adjacency topology feature extractor based on the convolutional neural network model to obtain a training semantic adjacency topology feature matrix; a training data semantic association unit 260, configured to pass the training service requirement description sentence granularity semantic coding feature vector sequence and the training semantic adjacency topology feature matrix through the inter-sentence semantic association encoder based on the graph convolution neural network model to obtain a training service requirement description inter-sentence semantic association coding feature vector sequence; the training data feature selection unit 270 is configured to pass the sequence of the training service requirement description inter-sentence semantic association coding feature vectors through the semantic gating-based essential feature attention selection network to obtain a training service requirement semantic understanding feature vector; a classification loss unit 280, configured to pass the training service requirement semantic understanding feature vector through the classifier-based flow basic structure recommender to obtain a classification loss function value; a model training unit 290, configured to train the semantic encoder based on the Transformer model, the semantic adjacency topological feature extractor based on the convolutional neural network model, the inter-sentence semantic association encoder based on the graph convolutional neural network model, the semantic gating-based essential feature attention selection network, and the classifier-based flow basic structure recommender with the classification loss function value, where in each iteration of the training, the training business requirement semantic understanding feature vector is optimized.
In the technical scheme of the application, each training service demand description inter-sentence semantic association coding feature vector in the sequence of the training service demand description inter-sentence semantic association coding feature vector respectively represents topological association coding features based on inter-sentence adjacent topological features for the semantic features of each service demand description sentence. When the sequence of the semantic association coding feature vectors among the training service demand description sentences passes through a semantic gating-based essential feature attention selection network, the semantic gating-based essential feature attention selection network screens important parts and unimportant parts in the sequence of the semantic association coding feature vectors among the training service demand description sentences based on a scoring mechanism, and filters the unimportant parts through a gating mechanism, so that the training service demand semantic understanding feature vectors have relatively better feature significance and important semantic information focusing, and meanwhile, the discretized local feature distribution of the training service demand semantic understanding feature vectors is also caused, and the convergence effect of the training service demand semantic understanding feature vectors to a class probability density space is affected when the training service demand semantic understanding feature vectors are classified through a flow basic structure recommender based on a classifier.
Therefore, in the technical scheme of the application, when the training service demand semantic understanding feature vector is classified and iterated through the classifier-based flow basic structure recommender, the training service demand semantic understanding feature vector is optimized according to the following optimization formula, wherein the optimization formula is as follows:
Wherein, Is the training business requirement semantic understanding feature vector,Is the semantic understanding feature vector of the training service requirementThe class probability values obtained by the classifier,Is the semantic understanding feature vector of the training service requirementIs the first of (2)The value of the characteristic is a value of,Is the semantic understanding feature vector of the training service requirementIs used for the average value of all the characteristic values of (a),Representing the operation of a natural exponential function,Represents the 1-norm of the feature vector, anIs the weight of the parameter to be exceeded,Is the first training business requirement semantic understanding feature vector after optimizationAnd characteristic values.
Specifically, the feature vector is understood semantically at the training business requirement
While maintaining the geometric details of the high-dimensional feature manifold under class probability mapping, understanding the feature vector of the feature value self details relative to the training service demand semantics by means of distortion consultation of class probabilities
Shape attribute editing is performed on the variant representation of the feature set of the training service requirement semantic understanding feature vector
The class probability distortion of (2) is mapped to the potential class space feature representation, and then consultation fusion is carried out by supplementing basic low-rank constraint potential inversion representation so as to make up the gap between the edited consultation representation and the original geometric shape detail, thereby improving the semantic understanding feature vector of the training service requirementCategory representation of the high-dimensional features of (c) to improve accuracy of the classification results.
In summary, the flow configuration system applied to the development operation and maintenance management platform according to the embodiment of the application is clarified, semantic analysis based on sentence granularity is performed on business requirement information input by a user by using artificial intelligence technology based on deep learning, sentence granularity semantic features of the business requirement information are captured, and inter-sentence semantic association coding is performed on the business requirement information based on inter-sentence semantic adjacency relations of the business requirement information, so that global semantic features of the business requirement information are fully extracted, and a proper flow basic structure is intelligently recommended according to the business requirement of the user. Therefore, the operation and maintenance efficiency is effectively improved, the operation and maintenance cost is reduced, and a more intelligent and efficient development operation and maintenance management platform is provided for enterprises.
Fig. 6 is a flowchart of a flow configuration method applied to an development operation and maintenance management platform according to an embodiment of the present application. As shown in fig. 6, a flow configuration method applied to an development operation and maintenance management platform according to an embodiment of the present application includes the steps of: s1, acquiring text description of service requirements input by a user; s2, carrying out sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of sentence-granularity semantic coding feature vectors of the service requirement description; s3, carrying out inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic coding feature vectors to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors; s4, extracting essential characteristics of the sequence of the semantic association coding feature vectors among the business requirement description sentences to obtain business requirement semantic understanding feature vectors; s5, determining the recommendation type of the basic flow structure based on the service demand semantic understanding feature vector. Here, it will be understood by those skilled in the art that the specific operations of the respective steps in the above-described flow configuration method applied to the development of the operation and maintenance management platform have been described in detail in the above description of the flow configuration system applied to the development of the operation and maintenance management platform with reference to fig. 1 to 5, and thus, repetitive descriptions thereof will be omitted. The basic principles of the present application have been described above in connection with specific embodiments, but it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not intended to be limiting, and these advantages, benefits, effects, etc. are not to be construed as necessarily possessed by the various embodiments of the application. Furthermore, the particular details of the above-described embodiments are for purposes of illustration and understanding only, and are not intended to limit the application to the particular details described above, but are not necessarily employed. In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments. In the several embodiments provided by the present application, it should be understood that the disclosed systems and methods may be implemented in other ways. For example, the system embodiments described above are merely illustrative, and for example, the module division is merely a logical function division, and other manners of division may be implemented in practice. The modules illustrated as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, each functional module in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units can be realized in a form of hardware or a form of hardware and a form of software functional modules.

Claims (3)

1. A process configuration system for an development operation and maintenance management platform, comprising:
the business requirement acquisition module is used for acquiring text description of business requirements input by a user;
the sentence-granularity semantic coding module is used for carrying out sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of the sentence-granularity semantic coding feature vectors of the service requirement description;
the inter-sentence semantic association coding module is used for carrying out inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic association coding feature vectors so as to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors;
The feature screening module is used for extracting essential features of the sequence of the semantic association coding feature vectors among the business requirement description sentences to obtain business requirement semantic understanding feature vectors;
the flow basic structure recommending module is used for determining the recommending type of the flow basic structure based on the service demand semantic understanding feature vector;
The system also comprises a training module for training a semantic encoder based on a transducer model, a semantic adjacency topological feature extractor based on a convolutional neural network model, an inter-sentence semantic association encoder based on a graph convolutional neural network model, a semantic gating-based essential feature attention selection network and a classifier-based flow basic structure recommender;
the training module comprises:
The training data acquisition unit is used for acquiring training data, wherein the training data comprises text description of training service requirements input by a user and a true value of a flow basic structure type;
The training data clause unit is used for carrying out clause processing on the text description of the training service requirement to obtain a sequence of training service requirement description sentences;
The training data semantic coding unit is used for respectively carrying out semantic coding on each training service demand description sentence in the sequence of the training service demand description sentences by using the semantic coder based on the Transformer model so as to obtain a sequence of training service demand description sentence granularity semantic coding feature vectors;
The training data semantic adjacency correlation unit is used for calculating a training semantic adjacency topology matrix of the sequence of the training business requirement description sentence granularity semantic coding feature vectors;
The training data semantic adjacency feature extraction unit is used for enabling the training semantic adjacency topological matrix to pass through the semantic adjacency topological feature extractor based on the convolutional neural network model so as to obtain a training semantic adjacency topological feature matrix;
The training data semantic association unit is used for enabling the training service requirement description sentence granularity semantic coding feature vector sequence and the training semantic adjacency topological feature matrix to pass through the inter-sentence semantic association encoder based on the graph convolution neural network model so as to obtain the training service requirement description inter-sentence semantic association coding feature vector sequence;
The training data feature selection unit is used for enabling the sequence of the training service requirement description inter-sentence semantic association coding feature vectors to pass through the essential feature attention selection network based on semantic gating to obtain training service requirement semantic understanding feature vectors;
The classification loss unit is used for passing the training service demand semantic understanding feature vector through the classifier-based flow basic structure recommender to obtain a classification loss function value;
A model training unit, configured to train the semantic encoder based on the Transformer model, the semantic adjacency topological feature extractor based on the convolutional neural network model, the inter-sentence semantic association encoder based on the graph convolutional neural network model, the semantic gating-based essential feature attention selection network, and the classifier-based flow basic structure recommender with the classification loss function value, where in each iteration of the training, the training business requirement semantic understanding feature vector is optimized;
And optimizing the training business demand semantic understanding feature vector by using the following optimization formula when the training business demand semantic understanding feature vector is classified and iterated by a flow basic structure recommender based on a classifier, wherein the optimization formula is as follows:
wherein, Is the training business requirement semantic understanding feature vector,Is the semantic understanding feature vector of the training service requirementThe class probability values obtained by the classifier,Is the semantic understanding feature vector of the training service requirementIs the first of (2)The value of the characteristic is a value of,Is the semantic understanding feature vector of the training service requirementIs used for the average value of all the characteristic values of (a),Representing the operation of a natural exponential function,Represents the 1-norm of the feature vector, anIs the weight of the parameter to be exceeded,Is the first training business requirement semantic understanding feature vector after optimizationA characteristic value;
the sentence-granularity semantic coding module comprises:
The clause processing unit is used for carrying out clause processing on the text description of the service requirement to obtain a sequence of the service requirement description sentence;
the semantic coding unit is used for respectively carrying out semantic coding on each business requirement description sentence in the sequence of the business requirement description sentences by using a semantic coder based on a Transformer model so as to obtain the sequence of the business requirement description sentence granularity semantic coding feature vector;
the inter-sentence semantic association coding module comprises:
The semantic adjacency topology matrix construction unit is used for calculating the semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors;
The semantic adjacency topological feature extraction unit is used for enabling the semantic adjacency topological matrix to pass through a semantic adjacency topological feature extractor based on a convolutional neural network model so as to obtain a semantic adjacency topological feature matrix;
the semantic association coding unit is used for enabling the sequence of the business requirement description sentence granularity semantic association coding feature vectors and the semantic adjacency topological feature matrix to pass through an inter-sentence semantic association coder based on a graph convolution neural network model so as to obtain the sequence of the business requirement description inter-sentence semantic association coding feature vectors;
the feature screening module is used for:
The sequence of the semantic association coding feature vectors among the business requirement description sentences is subjected to semantic gating-based essential feature attention selection network to obtain the business requirement semantic understanding feature vectors;
the feature screening module is used for:
Processing the sequence of the semantic association coding feature vectors among the business requirement description sentences by using the following essential feature attention selection formula to obtain the business requirement semantic understanding feature vectors; wherein, the essential feature attention selection formula is:
wherein, Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Representing the 1-norm of the feature vector,Describing the length-1 of the sequence of inter-sentence semantic association encoded feature vectors for the business requirement,A representation of a sequence of inter-sentence semantic association encoded feature vectors is described for the business requirement,The attention is given a score by a factor,The masking process is represented by a process of masking,Scoring coefficients for masking the attention,Representing the operation of a natural exponential function,Representing the total number of masked attention scoring coefficients,Semantic understanding of feature vectors for the business requirements;
the flow basic structure recommending module is used for:
And passing the service demand semantic understanding feature vector through a flow basic structure recommender based on a classifier to obtain a classification result, wherein the classification result is used for representing type labels of the flow basic structure.
2. The flow configuration system applied to the development operation and maintenance management platform according to claim 1, wherein the semantic adjacency topology matrix construction unit is configured to:
calculating a semantic adjacency topology matrix of the sequence of the business requirement description sentence granularity semantic coding feature vectors by using the following semantic adjacency association formula; the semantic adjacency association formula is as follows:
wherein, Is the first in the sequence of the business requirement description sentence granularity semantic coding feature vectorsThe individual business requirement descriptors granularity semantic coding feature vectors,Is the j-th business requirement description sentence granularity semantic coding feature vector,Is the firstStandard deviation between the granularity semantic coding feature vector of the individual business requirement description sentence and the granularity semantic coding feature vector of the jth business requirement description sentence,Representing the square of the 2 norms of the feature vectors,The element pair bit subtracting difference processing is represented,Representing the operation of a natural exponential function,For the first of the semantic adjacency topology matricesCharacteristic values of the location.
3. A flow configuration method for the flow configuration system applied to the development operation and maintenance platform according to claim 1, comprising:
acquiring text description of business requirements input by a user;
Performing sentence-granularity semantic coding on the text description of the service requirement to obtain a sequence of sentence-granularity semantic coding feature vectors of the service requirement description;
Performing inter-sentence semantic association coding on the sequence of the business requirement description sentence granularity semantic association coding feature vectors to obtain a sequence of business requirement description inter-sentence semantic association coding feature vectors;
extracting essential characteristics of the sequence of the semantic association coding feature vectors among the business requirement description sentences to obtain business requirement semantic understanding feature vectors;
Determining a recommendation type of a flow basic structure based on the service demand semantic understanding feature vector;
The method further comprises a training step of training a semantic encoder based on a transducer model, a semantic adjacency topological feature extractor based on a convolutional neural network model, an inter-sentence semantic association encoder based on a graph convolutional neural network model, a semantic gating-based essential feature attention selection network and a classifier-based flow basic structure recommender;
the training step comprises the following steps:
Acquiring training data, wherein the training data comprises text description of training service requirements input by a user and a true value of a flow basic structure type;
Sentence processing is carried out on the text description of the training service requirement to obtain a sequence of training service requirement description sentences;
Using the semantic encoder based on the transducer model to respectively carry out semantic encoding on each training service demand description sentence in the sequence of training service demand description sentences so as to obtain a sequence of training service demand description sentence granularity semantic encoding feature vectors;
Calculating a training semantic adjacency topology matrix of the sequence of the training service requirement description sentence granularity semantic coding feature vectors;
passing the training semantic adjacency topology matrix through the semantic adjacency topology feature extractor based on the convolutional neural network model to obtain a training semantic adjacency topology feature matrix;
the training business requirement description sentence granularity semantic coding feature vector sequence and the training semantic adjacency topological feature matrix are processed through the inter-sentence semantic association encoder based on the graph convolution neural network model to obtain the training business requirement description inter-sentence semantic association coding feature vector sequence;
the sequence of the semantic association coding feature vectors among the training service demand description sentences passes through the semantic gating-based essential feature attention selection network to obtain training service demand semantic understanding feature vectors;
Passing the training service demand semantic understanding feature vector through the classifier-based flow basic structure recommender to obtain a classification loss function value;
Training the semantic encoder based on the transducer model, the semantic adjacency topological feature extractor based on the convolutional neural network model, the inter-sentence semantic association encoder based on the graph convolutional neural network model, the semantic gating-based essential feature attention selection network and the classifier-based flow basic structure recommender by using the classification loss function values, wherein in each iteration of the training, the training business requirement semantic understanding feature vector is optimized;
And optimizing the training business demand semantic understanding feature vector by using the following optimization formula when the training business demand semantic understanding feature vector is classified and iterated by a flow basic structure recommender based on a classifier, wherein the optimization formula is as follows:
wherein, Is the training business requirement semantic understanding feature vector,Is the semantic understanding feature vector of the training service requirementThe class probability values obtained by the classifier,Is the semantic understanding feature vector of the training service requirementIs the first of (2)The value of the characteristic is a value of,Is the semantic understanding feature vector of the training service requirementIs used for the average value of all the characteristic values of (a),Representing the operation of a natural exponential function,Represents the 1-norm of the feature vector, anIs the weight of the parameter to be exceeded,Is the first training business requirement semantic understanding feature vector after optimizationA characteristic value;
The method further comprises the steps of:
The sequence of the semantic association coding feature vectors among the business requirement description sentences is subjected to semantic gating-based essential feature attention selection network to obtain the business requirement semantic understanding feature vectors;
Processing the sequence of the semantic association coding feature vectors among the business requirement description sentences by using the following essential feature attention selection formula to obtain the business requirement semantic understanding feature vectors; wherein, the essential feature attention selection formula is:
wherein, Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Is the first in the sequence of the semantic association coding feature vectors among the business requirement description sentencesThe individual business requirements describe inter-sentence semantic association coding feature vectors,Representing the 1-norm of the feature vector,Describing the length-1 of the sequence of inter-sentence semantic association encoded feature vectors for the business requirement,A representation of a sequence of inter-sentence semantic association encoded feature vectors is described for the business requirement,The attention is given a score by a factor,The masking process is represented by a process of masking,Scoring coefficients for masking the attention,Representing the operation of a natural exponential function,Representing the total number of masked attention scoring coefficients,Semantic understanding of feature vectors for the business requirements;
The method further comprises the steps of:
And passing the service demand semantic understanding feature vector through a flow basic structure recommender based on a classifier to obtain a classification result, wherein the classification result is used for representing type labels of the flow basic structure.
CN202410565476.3A 2024-05-09 2024-05-09 Flow configuration system and method applied to development operation and maintenance management platform Active CN118154135B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410565476.3A CN118154135B (en) 2024-05-09 2024-05-09 Flow configuration system and method applied to development operation and maintenance management platform

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410565476.3A CN118154135B (en) 2024-05-09 2024-05-09 Flow configuration system and method applied to development operation and maintenance management platform

Publications (2)

Publication Number Publication Date
CN118154135A CN118154135A (en) 2024-06-07
CN118154135B true CN118154135B (en) 2024-08-13

Family

ID=91298870

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410565476.3A Active CN118154135B (en) 2024-05-09 2024-05-09 Flow configuration system and method applied to development operation and maintenance management platform

Country Status (1)

Country Link
CN (1) CN118154135B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118553242B (en) * 2024-05-30 2025-04-01 浙江河马管家网络科技有限公司 Intelligent task navigation system and method based on machine learning
CN118535142B (en) * 2024-07-01 2024-10-22 武汉智网云达技术有限公司 Intelligent assisted coding system and method based on large model and machine learning
CN118710394A (en) * 2024-07-02 2024-09-27 日照财金路演有限公司 Financial debt early warning monitoring system and method based on artificial intelligence
CN119337380B (en) * 2024-09-29 2025-09-30 国网河南省电力公司信息通信分公司 Vulnerability Verification System Based on Dynamic Business Component Analysis
CN119004486B (en) * 2024-10-18 2025-02-25 北京关键科技股份有限公司 Intelligent screening method for software information of motor vehicle inspection and testing equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114492444A (en) * 2022-02-10 2022-05-13 北京工业大学 Chinese electronic medical case medical entity part-of-speech tagging method
CN117349437A (en) * 2023-10-07 2024-01-05 广东信拓人工智能研究院有限公司 Government information management system and method based on intelligent AI
CN117575865A (en) * 2024-01-17 2024-02-20 江西交通职业技术学院 College scientific research informationized management platform and method

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9400835B2 (en) * 2011-07-28 2016-07-26 Nokia Technologies Oy Weighting metric for visual search of entity-relationship databases
CN115146488B (en) * 2022-09-05 2022-11-22 山东鼹鼠人才知果数据科技有限公司 Variable business process intelligent modeling system and method based on big data
CN117560201A (en) * 2023-11-20 2024-02-13 临高里缲智能科技有限公司 Computer network information safety protection system and method
CN117993002A (en) * 2023-12-22 2024-05-07 国网河南省电力公司信息通信分公司 A data security protection method based on artificial intelligence
CN117710166A (en) * 2024-02-06 2024-03-15 吉林大学 Teaching guidance and detection system and method for pediatric neurological care
CN117992953A (en) * 2024-03-12 2024-05-07 国网河南省电力公司信息通信分公司 Abnormal user behavior identification method based on operation behavior tracking

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114492444A (en) * 2022-02-10 2022-05-13 北京工业大学 Chinese electronic medical case medical entity part-of-speech tagging method
CN117349437A (en) * 2023-10-07 2024-01-05 广东信拓人工智能研究院有限公司 Government information management system and method based on intelligent AI
CN117575865A (en) * 2024-01-17 2024-02-20 江西交通职业技术学院 College scientific research informationized management platform and method

Also Published As

Publication number Publication date
CN118154135A (en) 2024-06-07

Similar Documents

Publication Publication Date Title
CN118154135B (en) Flow configuration system and method applied to development operation and maintenance management platform
CN110597735B (en) A software defect prediction method for deep learning of open source software defect features
CN113177124B (en) Method and system for constructing knowledge graph in vertical field
CN106096727B (en) A kind of network model building method and device based on machine learning
CN117312989A (en) Context-aware column semantic recognition method and system based on GCN and RoBERTa
CN112597296A (en) Abstract generation method based on plan mechanism and knowledge graph guidance
Lin et al. Deep structured scene parsing by learning with image descriptions
CN118886336A (en) A method for constructing railway engineering design knowledge base based on cloud platform
CN116150404A (en) A multi-modal knowledge map construction method for educational resources based on federated learning
CN115470786A (en) Entity Information Extraction Method for Power Defect Text Based on Improved Transformer Encoder
CN118012990A (en) A prompt text generation method, system, computer device and storage medium
CN118964594A (en) Power grid standard document analysis and information extraction method based on AI
CN119475232A (en) Public opinion analysis method, device and medium based on big data and deep learning
CN116541166A (en) Super-computing power scheduling server and resource management method
CN119669476A (en) A short text clustering method and device based on large language model
CN113656560B (en) Emotion category prediction method and device, storage medium and electronic equipment
CN115757762A (en) A Text Summarization Method Combining Generative and Extractive
Yu et al. Workflow recommendation based on graph embedding
CN118295698A (en) Design document generation method and device, electronic equipment and storage medium
CN115964953A (en) Power grid digital resource modeling management method based on meta-learning
CN116563679A (en) Abnormal multimedia resource identification method and device, electronic equipment and storage medium
Zhang et al. Predicting Relations in SG-CIM Model Based on Graph Structure and Semantic Information
CN113723078A (en) Text logic information structuring method and device and electronic equipment
CN114385892A (en) Article grade identification method and device, server and storage medium
CN109299442A (en) Method and System for Recognition of Primary and Secondary Relationships in Chinese Texts

Legal Events

Date Code Title Description
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant