CN117610658A - Knowledge graph data dynamic updating method and system based on artificial intelligence - Google Patents
Knowledge graph data dynamic updating method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN117610658A CN117610658A CN202311769481.8A CN202311769481A CN117610658A CN 117610658 A CN117610658 A CN 117610658A CN 202311769481 A CN202311769481 A CN 202311769481A CN 117610658 A CN117610658 A CN 117610658A
- Authority
- CN
- China
- Prior art keywords
- entity
- semantic
- text description
- sequence
- granularity
- 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.)
- Pending
Links
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 44
- 239000013598 vector Substances 0.000 claims abstract description 194
- 230000004927 fusion Effects 0.000 claims abstract description 92
- 238000004458 analytical method Methods 0.000 claims abstract description 39
- 230000003993 interaction Effects 0.000 claims abstract description 34
- 238000007499 fusion processing Methods 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims description 19
- 230000011218 segmentation Effects 0.000 claims description 19
- 238000012937 correction Methods 0.000 claims description 8
- 238000011176 pooling Methods 0.000 claims description 8
- 238000005457 optimization Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 6
- 230000014509 gene expression Effects 0.000 description 6
- 238000000605 extraction Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000008451 emotion Effects 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 238000013519 translation Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a knowledge graph data dynamic updating method and system based on artificial intelligence, and relates to the field of data dynamic updating. Firstly, carrying out semantic analysis based on word granularity on current text description of a first entity to obtain a sequence of semantic feature vectors of the first entity current text description word granularity, then carrying out semantic analysis based on word granularity on updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description word granularity, then carrying out semantic interaction fusion processing on the sequence of semantic feature vectors of the first entity current text description word granularity and the sequence of semantic feature vectors of the first entity updated text description word granularity to obtain first entity fusion semantic features, and finally, generating optimized updated text description of the first entity based on the first entity fusion semantic features. Thus, the dynamic updating of the knowledge graph data can be realized, and the accuracy of the knowledge graph is improved.
Description
Technical Field
The present application relates to the field of dynamic data updating, and more particularly, to a method and system for dynamically updating knowledge-graph data based on artificial intelligence.
Background
Knowledge graph is a technique for representing and storing structured and semi-structured data that organizes entities, attributes, and relationships into a directed graph, thereby providing an efficient method of data querying and analysis. However, as the data changes continuously, the data in the knowledge-graph may be outdated or incomplete, resulting in degradation of the quality of the knowledge-graph. Therefore, dynamic updating of the knowledge-graph data is required to maintain synchronization with the real world.
However, the existing knowledge graph data updating method mainly depends on manual editing or a rule-based strategy, and the method needs a great deal of time and effort to search and update the knowledge graph data manually, lacks automation capability, limits the instantaneity and expansibility of the knowledge graph, and can cause the slow updating speed of the knowledge graph and not reflect the real world change in time. In addition, because the knowledge graph data has wide and various information sources, the existing scheme has difficulty in processing large-scale knowledge graphs and mass data, so that the traditional manual or semi-automatic knowledge graph data updating method cannot meet the requirements of efficient processing and updating.
Accordingly, a dynamic update scheme for knowledge-graph data based on artificial intelligence is desired.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a knowledge graph data dynamic updating method and system based on artificial intelligence, which can realize dynamic updating of knowledge graph data, thereby improving accuracy and practicability of the knowledge graph.
According to one aspect of the present application, there is provided a knowledge-graph data dynamic updating method based on artificial intelligence, which includes:
extracting a current text description of the first entity from the knowledge-graph;
extracting an updated text description of the first entity from the internet;
performing semantic analysis based on word granularity on the current text description of the first entity to obtain a sequence of semantic feature vectors of the word granularity of the current text description of the first entity;
performing semantic analysis based on word granularity on the updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description word granularity;
carrying out semantic interaction fusion processing on the sequence of the semantic feature vector with the granularity of the current text description word of the first entity and the sequence of the semantic feature vector with the granularity of the updated text description word of the first entity so as to obtain a first entity fusion semantic feature; and
and generating an optimized updating text description of the first entity based on the first entity fusion semantic features.
According to another aspect of the present application, there is provided an artificial intelligence based knowledge-graph data dynamic update system, which includes:
the current data acquisition module is used for extracting the current text description of the first entity from the knowledge graph;
the update data acquisition module is used for extracting an update text description of the first entity from the Internet;
the current text semantic analysis module is used for carrying out semantic analysis based on word granularity on the current text description of the first entity so as to obtain a sequence of semantic feature vectors with the word granularity of the current text description of the first entity;
the update text semantic analysis module is used for carrying out semantic analysis based on word granularity on the update text description of the first entity to obtain a sequence of semantic feature vectors of the first entity update text description word granularity;
the semantic interaction fusion module is used for carrying out semantic interaction fusion processing on the sequence of the semantic feature vector with the granularity of the current text description word of the first entity and the sequence of the semantic feature vector with the granularity of the updated text description word of the first entity so as to obtain a first entity fusion semantic feature; and
and the generating module is used for generating the optimized updating text description of the first entity based on the fusion semantic features of the first entity.
Compared with the prior art, the method and the system for dynamically updating the knowledge graph data based on the artificial intelligence are characterized in that firstly, semantic analysis based on word granularity is carried out on current text description of a first entity to obtain a sequence of semantic feature vectors of the first entity current text description word granularity, then, semantic analysis based on word granularity is carried out on updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description word granularity, then, semantic interaction fusion processing is carried out on the sequence of the first entity current text description word granularity semantic feature vectors and the sequence of semantic feature vectors of the first entity updated text description word granularity to obtain first entity fusion semantic features, and finally, the optimized updated text description of the first entity is generated based on the first entity fusion semantic features. Thus, the dynamic updating of the knowledge graph data can be realized, and the accuracy of the knowledge graph is improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly introduced below, which are not intended to be drawn to scale in terms of actual dimensions, with emphasis on illustrating the gist of the present application.
Fig. 1 is a flowchart of a method for dynamically updating knowledge-graph data based on artificial intelligence according to an embodiment of the application.
Fig. 2 is a schematic architecture diagram of an artificial intelligence-based knowledge-graph data dynamic updating method according to an embodiment of the application.
Fig. 3 is a flowchart of substep S160 of the artificial intelligence based knowledge-graph data dynamic update method, in accordance with an embodiment of the present application.
FIG. 4 is a block diagram of an artificial intelligence based knowledge-graph data dynamic update system, in accordance with an embodiment of the application.
Fig. 5 is an application scenario diagram of an artificial intelligence-based knowledge-graph data dynamic update method according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some, but not all embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present application without making any inventive effort, are also within the scope of the present application.
As used in this application and in the claims, the terms "a," "an," "the," and/or "the" are not specific to the singular, but may include the plural, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
Flowcharts are used in this application to describe the operations performed by systems according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Aiming at the technical problems, an artificial intelligence-based knowledge graph data dynamic updating method is provided, the method uses the Internet as a data source, extracts the current text description of a first entity from a knowledge graph, extracts the updated text description of the first entity from the Internet, and introduces an artificial intelligence-based semantic understanding algorithm at the rear end to perform semantic analysis and fusion of the current text description and the updated text description, thereby realizing automatic generation of the optimized updated text description of the first entity in the knowledge graph. Therefore, the dynamic updating of the knowledge graph data can be realized, and the accuracy and the practicability of the knowledge graph are improved.
Fig. 1 is a flowchart of a method for dynamically updating knowledge-graph data based on artificial intelligence according to an embodiment of the application.
Fig. 2 is a schematic architecture diagram of an artificial intelligence-based knowledge-graph data dynamic updating method according to an embodiment of the application. As shown in fig. 1 and fig. 2, the method for dynamically updating knowledge-graph data based on artificial intelligence according to an embodiment of the present application includes the steps of: s110, extracting a current text description of the first entity from the knowledge graph; s120, extracting updated text description of the first entity from the Internet; s130, carrying out semantic analysis based on word granularity on the current text description of the first entity to obtain a sequence of semantic feature vectors with the word granularity of the current text description of the first entity; s140, carrying out semantic analysis based on word granularity on the updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description word granularity; s150, carrying out semantic interaction fusion processing on the sequence of the semantic feature vector of the granularity of the current text description word of the first entity and the sequence of the semantic feature vector of the granularity of the updated text description word of the first entity so as to obtain a first entity fusion semantic feature; and S160, generating an optimized update text description of the first entity based on the first entity fusion semantic features.
Specifically, in the technical scheme of the application, first, a current text description of a first entity is extracted from a knowledge graph, and an updated text description of the first entity is extracted from the internet. Next, for a current text description of the first entity, considering that there is a lot of semantic information in the current text description, wherein the current text description is composed of a plurality of words, each word has associated semantics with each other. Therefore, in order to perform semantic analysis and understanding on the current text description, in the technical scheme of the application, after word segmentation processing is performed on the current text description of the first entity, semantic coding is performed in a current text description semantic encoder comprising a word embedding layer, so that semantic association characteristic information of the current text description of the first entity based on word granularity is extracted, and therefore a sequence of first entity current text description word granularity semantic characteristic vectors is obtained.
Accordingly, in step S130, performing semantic analysis based on word granularity on the current text description of the first entity to obtain a sequence of semantic feature vectors of the word granularity of the current text description of the first entity, including: and after word segmentation processing is carried out on the current text description of the first entity, a current text description semantic encoder comprising a word embedding layer is used for obtaining a sequence of the first entity current text description word granularity semantic feature vector.
Similarly, in order to enable semantic analysis and understanding of the updated text description of the first entity, in the technical solution of the present application, after word segmentation processing is performed on the updated text description of the first entity, semantic encoding is performed in an updated text description semantic encoder including a word embedding layer, so as to extract semantic association feature information of the updated text description based on word granularity, thereby obtaining a sequence of word granularity semantic feature vectors of the updated text description of the first entity.
Accordingly, in step S140, performing semantic analysis based on word granularity on the updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description word granularity, including: after word segmentation is carried out on the updated text description of the first entity, a sequence of the first entity updated text description word granularity semantic feature vector is obtained through an updated text description semantic encoder comprising a word embedding layer.
It should be appreciated that word segmentation is an important task in natural language processing that segments a continuous sequence of text into individual words or words, and in the above description word segmentation is used to process the current and updated text descriptions of the first entity. The purpose of word segmentation is to divide a continuous sequence of text into discrete words so that a computer can understand and process the text. The following are several important uses of word segmentation: 1. semantic understanding: the segmentation may divide the text into smaller semantic units so that the computer can understand and process the meaning of each word, which is important for subsequent semantic analysis and understanding tasks. 2. Feature extraction: the segmentation may convert text into a sequence of words, each of which may be considered a feature that may be used to construct a representation vector of the text, and thus for training and application of machine learning and deep learning models. 3. Information retrieval: the word segmentation can help a search engine and an information retrieval system to more accurately match the content of a user query and a document, and the accuracy and recall rate of retrieval can be improved by carrying out word segmentation on both the query and the document. Generally speaking, word segmentation processes convert continuous text sequences into discrete word sequences, providing the basis for subsequent semantic analysis, feature extraction, and text processing tasks.
The word embedding layer (Word Embedding Layer) is a layer for representing discrete words as vectors in a continuous vector space, which maps each word to a low-dimensional real vector, so that semantic relationships between words can be captured and represented in the vector space. The word embedding layer is used for converting discrete words into continuous vector representation, so that the problem of a traditional single-hot coding representation mode is solved. The traditional single-hot coding representation represents each word as a high-dimensional sparse vector, where only one element is 1 and the remaining elements are 0. This representation does not directly capture semantic similarity and relevance between terms. And the word embedding layer maps words into a continuous vector space by learning semantic relationships between words. In this vector space, the distance and direction between words may represent the semantic relationship between them. For example, semantically similar terms are closer together in vector space, while semantically unrelated terms are farther apart in vector space. The word embedding layer has a wide range of applications including, but not limited to, the following: semantic representation and semantic similarity calculation: semantic similarity between words can be measured by calculating the similarity between word vectors; text classification and emotion analysis: the word vector can be used as input to carry out text classification and emotion analysis tasks; machine translation and language generation: the word vector can be used as input for machine translation and language generation tasks; named entity identification and relationship extraction: the term vectors may be used to represent entities and relationships for named entity recognition and relationship extraction tasks. In other words, the word embedding layer provides an efficient way to represent and process natural language text by mapping discrete words into a continuous vector space, which plays an important role in natural language processing tasks.
It should be appreciated that, since the sequence of the first entity's current text descriptor granularity semantic feature vector contains word granularity-based semantic feature information in the current text description, and the sequence of the first entity's updated text descriptor granularity semantic feature vector contains word granularity-based semantic feature information in the updated text description. Also, it is contemplated that the current text description and the updated text description may contain different semantic information, e.g., the current text description may reflect basic attributes of the first entity, while the updated text description may provide new related information. Therefore, in order to fuse the semantic feature vectors of the two to facilitate comprehensive utilization of the information to obtain more comprehensive and accurate first entity description, in the technical scheme of the application, a semantic interaction fusion module is further used for processing the sequence of the semantic feature vector with the granularity of the current text description word of the first entity and the sequence of the semantic feature vector with the granularity of the updated text description word of the first entity to obtain the first entity fusion semantic feature vector. In particular, by using the semantic interaction fusion module to carry out semantic interaction fusion based on word granularity between the current text description and the updated text description, the semantic information can be promoted to be mutually supplemented and associated, so that richer and more accurate first entity semantic features are obtained, and more accurate and comprehensive feature input is provided for subsequent entity description text generation.
Accordingly, in step S150, performing semantic interaction fusion processing on the sequence of the semantic feature vector of the granularity of the current text descriptor of the first entity and the sequence of the semantic feature vector of the granularity of the updated text descriptor of the first entity to obtain a first entity fusion semantic feature, including: and processing the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector by using a semantic interaction fusion module to obtain a first entity fusion semantic feature vector as the first entity fusion semantic feature.
It should be understood that the function of the semantic interaction fusion module is to process the current text descriptor granularity semantic feature vector sequence and the updated text descriptor granularity semantic feature vector sequence of the first entity to obtain the fused semantic feature of the first entity, and the purpose of the module is to interact and fuse the semantic information in the two sequences, so as to obtain a richer and more comprehensive semantic representation. The main uses of the semantic interaction fusion module include the following aspects: 1. semantic fusion: by fusing the semantic feature vectors of the current text description and the updated text description, semantic information of the current text description and the updated text description can be mutually supplemented and fused, and more comprehensive and accurate semantic representation can be obtained. 2. Context understanding: the semantic interaction fusion module may help understand the contextual relationship between the current textual description and the updated textual description. Through interaction and fusion of semantic features of the two sequences, relevance and change between the two sequences can be captured, so that semantic meaning of the text can be better understood. 3. Feature enhancement: the semantic interaction fusion module can enhance semantic features of the text description, so that the representation is richer and the expression is more accurate. By fusing semantic information of different sources, more context information and semantic association can be provided, and the performance of subsequent tasks can be improved. 4. Semantic consistency: the semantic interaction fusion module can help to maintain semantic consistency between the current text description and the updated text description, and semantic representations of two sequences can be made to be closer in vector space through interaction fusion, so that semantic relevance of the two sequences is maintained. In summary, the semantic interaction fusion module realizes interaction and fusion of semantic information by processing the current text description and updating the semantic feature vector sequence of the text description, provides more comprehensive and richer semantic representation, and plays an important role in improving semantic understanding, context understanding, feature enhancement and the like.
Specifically, using a semantic interaction fusion module to process the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector to obtain a first entity fusion semantic feature vector as the first entity fusion semantic feature, including: performing attention enhancement based on the correlation degree between the sequence of the first entity current text descriptor granularity semantic feature vectors and the sequence of the first entity updated text descriptor granularity semantic feature vectors to obtain a sequence of attention enhanced first entity current text descriptor granularity semantic feature vectors and a sequence of attention enhanced first entity updated text descriptor granularity semantic feature vectors; fusing the sequence of the first entity current text descriptor granularity semantic feature vectors and feature vectors of corresponding positions in the sequence of the attention-enhanced first entity current text descriptor granularity semantic feature vectors to obtain a sequence of first entity current text description local fusion feature vectors, and fusing the sequence of the first entity update text descriptor granularity semantic feature vectors and feature vectors of corresponding positions in the sequence of the attention-enhanced first entity update text descriptor granularity semantic feature vectors to obtain a sequence of first entity update text description local fusion feature vectors; performing maximum value pooling processing on the sequence of the first entity current text description local fusion feature vector to obtain a first entity current text description local fusion maximum value pooling feature vector, and performing maximum value pooling processing on the sequence of the first entity update text description local fusion feature vector to obtain a first entity update text description local fusion maximum value pooling feature vector; and fusing the local fusion maximum value pooled feature vector of the current text description of the first entity and the local fusion maximum value pooled feature vector of the updated text description of the first entity to obtain the first entity fusion semantic feature vector.
And then, the semantic feature vector fused by the first entity passes through an entity description text generator based on an AIGC model to obtain the optimized updated text description of the first entity. That is, by inputting the fused semantic feature vector of the first entity into the AIGC model, the generation capability of the model can be utilized to integrate the semantics of the current text description and the semantics of the updated text description to generate an optimized updated text description, thereby improving the quality of the text description. This helps to improve the accuracy and practicality of knowledge-graph data.
Accordingly, in step S160, as shown in fig. 3, generating an optimized update text description of the first entity based on the first entity fusion semantic feature includes: s161, performing feature distribution optimization on the first entity fusion semantic feature vector to obtain an optimized first entity fusion semantic feature vector; and S162, enabling the optimized first entity fusion semantic feature vector to pass through an AIGC model-based entity description text generator to obtain an optimized updated text description of the first entity.
It should be understood that, in step S161, the quality and the expressive power of the semantic features may be further improved by performing feature distribution optimization on the first entity fusion semantic feature vector, where the objective of feature distribution optimization is to adjust the weights of the dimensions in the feature vector, so that each dimension can better capture relevant semantic information, which may be implemented by various optimization algorithms and technologies, such as a dimension reduction algorithm, a regularization method, a clustering algorithm, and so on. The optimized first entity fusion semantic feature vector can better reflect the semantic features of the first entity, and the quality and accuracy of text description generated in the subsequent steps are improved. In step S162, an optimized update text description of the first entity is generated using the optimized first entity fusion semantic feature vector by an AIGC (Attention-based Image Captioning) model-based entity description text generator. The AIGC model is an image description generation model based on an attention mechanism, which can generate a natural language description related to image contents by learning the correlation between images and texts. Inputting the optimized first entity fusion semantic feature vector into an AIGC model, and generating a corresponding optimized updating text description by the model according to the information of the feature vector. The generated textual description may more accurately reflect the characteristics and status of the first entity, providing more rich and useful information. In combination, the purpose of steps S161 and S162 is to improve the description and understanding capabilities of the first entity by optimizing the first entity to fuse semantic feature vectors and generating optimized updated text descriptions, providing more accurate and useful information for subsequent applications and analyses.
In particular, in the above technical solution, the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector are used to express the current text description of the first entity and the source semantic context association encoding text semantic feature based on word granularity of the update text description of the first entity, respectively, so that when the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector are processed by using the semantic interaction fusion module, the source semantic context association difference between the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector may cause semantic interaction sparsity between the feature sequences of semantic features, thereby affecting the expression effect of the first entity fusion semantic feature vector, and therefore, it is desirable to improve the semantic interaction sparsity expression of the first entity based on the respective feature expressions of the sequence of the first entity current descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector, thereby improving the semantic interaction sparsity expression of the first entity.
Based on this, the applicant of the present application corrects the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector.
Accordingly, in step S161, performing feature distribution optimization on the first entity fusion semantic feature vector to obtain an optimized first entity fusion semantic feature vector, including: correcting the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity updated text descriptor granularity semantic feature vector to obtain a corrected feature vector; and fusing the correction feature vector and the first entity fusion semantic feature vector to obtain the optimized first entity fusion semantic feature vector.
Specifically, correcting the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity updated text descriptor granularity semantic feature vector to obtain corrected feature vectors, including: correcting the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity updated text descriptor granularity semantic feature vector by using the following correction formula to obtain the correction feature vector; wherein, the correction formula is:
wherein V is 1 Is a first cascade feature vector obtained by cascading the sequence of the semantic feature vectors of the granularity of the current text description word of the first entity, and V 2 Is a second cascade feature vector obtained by cascading the sequence of the text descriptor granularity semantic feature vectors of the first entity update,representing the position-wise evolution of feature vectors, v 1max -1 And v 2max -1 Respectively the feature vectors V 1 And V 2 Maximum specialInverse of the sign value, alpha and beta are weight superparameters, ++indicates multiplication by location point, ++>Representing the subtraction of vectors, V c Is the correction feature vector.
Here, the pre-segmented local group of feature value sets is obtained by the sequence of the first entity current text descriptor granularity semantic feature vector and the evolution value of each feature value of the sequence of the first entity update text descriptor granularity semantic feature vector, and the key maximum feature of the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector is regressed therefrom, so that the per-position saliency distribution of feature values can be promoted based on the concept of the furthest point sampling, thereby performing sparse interactivity control among feature vectors by the key features with the saliency distribution, so as to realize correction of the feature vector V c And updating the restoration of the original feature manifold geometric representation of the sequence of text descriptor granularity semantic feature vectors for the first entity and the sequence of the first entity. Thus, the correction feature vector V is further used c And fusing the first entity fusion semantic feature vector with the first entity fusion semantic feature vector, so that the expression effect of the first entity fusion semantic feature vector can be improved, and the text description quality of the optimized update text description obtained by an AIGC model-based entity description text generator is improved. In this way, text description of the entity can be optimized and updated based on fusion semantics between the current text and the updated text of the entity in the knowledge graph, and by the mode, dynamic update of knowledge graph data can be realized, so that timeliness and accuracy of the knowledge graph are improved.
In summary, the method for dynamically updating the knowledge-graph data based on the artificial intelligence according to the embodiment of the application is explained, and the method can dynamically update the knowledge-graph data, so that the accuracy and the practicability of the knowledge-graph are improved.
Fig. 4 is a block diagram of an artificial intelligence based knowledge-graph data dynamic update system 100, in accordance with an embodiment of the application. As shown in fig. 4, the artificial intelligence based knowledge-graph data dynamic update system 100 according to an embodiment of the present application includes: a current data acquisition module 110, configured to extract a current text description of the first entity from the knowledge-graph; an update data acquisition module 120, configured to extract an update text description of the first entity from the internet; the current text semantic analysis module 130 is configured to perform semantic analysis based on word granularity on the current text description of the first entity to obtain a sequence of semantic feature vectors with the word granularity of the current text description of the first entity; the update text semantic analysis module 140 is configured to perform semantic analysis based on word granularity on the update text description of the first entity to obtain a sequence of semantic feature vectors with the word granularity of the update text description of the first entity; the semantic interaction fusion module 150 is configured to perform semantic interaction fusion processing on the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector to obtain a first entity fusion semantic feature; and a generating module 160, configured to generate an optimized update text description of the first entity based on the first entity fusion semantic feature.
In one example, in the above-mentioned artificial intelligence based knowledge-graph data dynamic updating system 100, the current text semantic analysis module 130 is configured to: and after word segmentation processing is carried out on the current text description of the first entity, a current text description semantic encoder comprising a word embedding layer is used for obtaining a sequence of the first entity current text description word granularity semantic feature vector.
In one example, in the above-described artificial intelligence based knowledge-graph data dynamic update system 100, the update text semantic analysis module 140 is configured to: after word segmentation is carried out on the updated text description of the first entity, a sequence of the first entity updated text description word granularity semantic feature vector is obtained through an updated text description semantic encoder comprising a word embedding layer.
Here, it will be understood by those skilled in the art that the specific functions and operations of the respective modules in the above-described artificial intelligence-based knowledge-graph data dynamic updating system 100 have been described in detail in the above description of the artificial intelligence-based knowledge-graph data dynamic updating method with reference to fig. 1 to 3, and thus, repetitive descriptions thereof will be omitted.
As described above, the artificial intelligence based knowledge-graph data dynamic update system 100 according to the embodiment of the present application may be implemented in various wireless terminals, for example, a server or the like having an artificial intelligence based knowledge-graph data dynamic update algorithm. In one example, the artificial intelligence based knowledge-graph data dynamic update system 100 according to embodiments of the present application may be integrated into a wireless terminal as a software module and/or hardware module. For example, the artificial intelligence based knowledge-graph data dynamic update system 100 may be a software module in the operating system of the wireless terminal or may be an application developed for the wireless terminal; of course, the artificial intelligence based knowledge-graph data dynamic updating system 100 can also be one of a plurality of hardware modules of the wireless terminal.
Alternatively, in another example, the artificial intelligence based knowledge-graph data dynamic update system 100 and the wireless terminal may be separate devices, and the artificial intelligence based knowledge-graph data dynamic update system 100 may be connected to the wireless terminal through a wired and/or wireless network and transmit interaction information in a agreed data format.
Fig. 5 is an application scenario diagram of an artificial intelligence-based knowledge-graph data dynamic update method according to an embodiment of the present application. As shown in fig. 5, in the application scenario, first, a current text description of a first entity is extracted from a knowledge-graph (e.g., D1 illustrated in fig. 5), then, an updated text description of the first entity is extracted from the internet (e.g., D2 illustrated in fig. 5), and then, the current text description of the first entity and the updated text description of the first entity are input into a server (e.g., S illustrated in fig. 5) deployed with an artificial intelligence-based knowledge-graph data dynamic update algorithm, wherein the server is capable of processing the current text description of the first entity and the updated text description of the first entity using the artificial intelligence-based knowledge-graph data dynamic update algorithm to obtain an optimized updated text description of the first entity.
Furthermore, those skilled in the art will appreciate that the various aspects of the invention are illustrated and described in the context of a number of patentable categories or circumstances, including any novel and useful procedures, machines, products, or materials, or any novel and useful modifications thereof. Accordingly, aspects of the present application may be performed entirely by hardware, entirely by software (including firmware, resident software, micro-code, etc.) or by a combination of hardware and software. The above hardware or software may be referred to as a "data block," module, "" engine, "" unit, "" component, "or" system. Furthermore, aspects of the present application may take the form of a computer product, comprising computer-readable program code, embodied in one or more computer-readable media.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
The foregoing is illustrative of the present invention and is not to be construed as limiting thereof. Although a few exemplary embodiments of this invention have been described, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without materially departing from the novel teachings and advantages of this invention. Accordingly, all such modifications are intended to be included within the scope of this invention as defined in the following claims. It is to be understood that the foregoing is illustrative of the present invention and is not to be construed as limited to the specific embodiments disclosed, and that modifications to the disclosed embodiments, as well as other embodiments, are intended to be included within the scope of the appended claims. The invention is defined by the claims and their equivalents.
Claims (10)
1. The method for dynamically updating the knowledge graph data based on the artificial intelligence is characterized by comprising the following steps of:
extracting a current text description of the first entity from the knowledge-graph;
extracting an updated text description of the first entity from the internet;
performing semantic analysis based on word granularity on the current text description of the first entity to obtain a sequence of semantic feature vectors of the word granularity of the current text description of the first entity;
performing semantic analysis based on word granularity on the updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description word granularity;
carrying out semantic interaction fusion processing on the sequence of the semantic feature vector with the granularity of the current text description word of the first entity and the sequence of the semantic feature vector with the granularity of the updated text description word of the first entity so as to obtain a first entity fusion semantic feature; and
and generating an optimized updating text description of the first entity based on the first entity fusion semantic features.
2. The method for dynamically updating knowledge-graph data based on artificial intelligence of claim 1, wherein performing semantic analysis based on word granularity on the current text description of the first entity to obtain a sequence of semantic feature vectors of the first entity's current text description word granularity comprises:
and after word segmentation processing is carried out on the current text description of the first entity, a current text description semantic encoder comprising a word embedding layer is used for obtaining a sequence of the first entity current text description word granularity semantic feature vector.
3. The method for dynamically updating knowledge-graph data based on artificial intelligence of claim 2, wherein performing semantic analysis based on word granularity on the updated text description of the first entity to obtain a sequence of semantic feature vectors of the first entity updated text description granularity comprises:
after word segmentation is carried out on the updated text description of the first entity, a sequence of the first entity updated text description word granularity semantic feature vector is obtained through an updated text description semantic encoder comprising a word embedding layer.
4. The method for dynamically updating knowledge-graph data based on artificial intelligence according to claim 3, wherein performing semantic interaction fusion processing on the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity updated text descriptor granularity semantic feature vector to obtain a first entity fusion semantic feature comprises:
and processing the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity update text descriptor granularity semantic feature vector by using a semantic interaction fusion module to obtain a first entity fusion semantic feature vector as the first entity fusion semantic feature.
5. The method for dynamically updating knowledge-graph data based on artificial intelligence of claim 4, wherein processing the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity updated text descriptor granularity semantic feature vector using a semantic interaction fusion module to obtain a first entity fusion semantic feature vector as the first entity fusion semantic feature comprises:
performing attention enhancement based on the correlation degree between the sequence of the first entity current text descriptor granularity semantic feature vectors and the sequence of the first entity updated text descriptor granularity semantic feature vectors to obtain a sequence of attention enhanced first entity current text descriptor granularity semantic feature vectors and a sequence of attention enhanced first entity updated text descriptor granularity semantic feature vectors;
fusing the sequence of the first entity current text descriptor granularity semantic feature vectors and feature vectors of corresponding positions in the sequence of the attention-enhanced first entity current text descriptor granularity semantic feature vectors to obtain a sequence of first entity current text description local fusion feature vectors, and fusing the sequence of the first entity update text descriptor granularity semantic feature vectors and feature vectors of corresponding positions in the sequence of the attention-enhanced first entity update text descriptor granularity semantic feature vectors to obtain a sequence of first entity update text description local fusion feature vectors;
performing maximum value pooling processing on the sequence of the first entity current text description local fusion feature vector to obtain a first entity current text description local fusion maximum value pooling feature vector, and performing maximum value pooling processing on the sequence of the first entity update text description local fusion feature vector to obtain a first entity update text description local fusion maximum value pooling feature vector; and
and fusing the local fusion maximum value pooled feature vector of the current text description of the first entity and the local fusion maximum value pooled feature vector of the updated text description of the first entity to obtain the first entity fusion semantic feature vector.
6. The method for dynamically updating knowledge-graph data based on artificial intelligence of claim 5, wherein generating an optimized update text description of the first entity based on the first entity fusion semantic features comprises:
performing feature distribution optimization on the first entity fusion semantic feature vector to obtain an optimized first entity fusion semantic feature vector; and
and the optimized first entity fusion semantic feature vector passes through an AIGC model-based entity description text generator to obtain an optimized updated text description of the first entity.
7. The method for dynamically updating knowledge-graph data based on artificial intelligence of claim 6, wherein performing feature distribution optimization on the first entity fusion semantic feature vector to obtain an optimized first entity fusion semantic feature vector, comprises:
correcting the sequence of the first entity current text descriptor granularity semantic feature vector and the sequence of the first entity updated text descriptor granularity semantic feature vector to obtain a corrected feature vector; and
and fusing the correction feature vector and the first entity fusion semantic feature vector to obtain the optimized first entity fusion semantic feature vector.
8. A knowledge-graph data dynamic updating system based on artificial intelligence, comprising:
the current data acquisition module is used for extracting the current text description of the first entity from the knowledge graph;
the update data acquisition module is used for extracting an update text description of the first entity from the Internet;
the current text semantic analysis module is used for carrying out semantic analysis based on word granularity on the current text description of the first entity so as to obtain a sequence of semantic feature vectors with the word granularity of the current text description of the first entity;
the update text semantic analysis module is used for carrying out semantic analysis based on word granularity on the update text description of the first entity to obtain a sequence of semantic feature vectors of the first entity update text description word granularity;
the semantic interaction fusion module is used for carrying out semantic interaction fusion processing on the sequence of the semantic feature vector with the granularity of the current text description word of the first entity and the sequence of the semantic feature vector with the granularity of the updated text description word of the first entity so as to obtain a first entity fusion semantic feature; and
and the generating module is used for generating the optimized updating text description of the first entity based on the fusion semantic features of the first entity.
9. The artificial intelligence based knowledge-graph data dynamic updating system of claim 8, wherein the current text semantic analysis module is configured to:
and after word segmentation processing is carried out on the current text description of the first entity, a current text description semantic encoder comprising a word embedding layer is used for obtaining a sequence of the first entity current text description word granularity semantic feature vector.
10. The artificial intelligence based knowledge-graph data dynamic updating system of claim 9, wherein the updated text semantic analysis module is configured to:
after word segmentation is carried out on the updated text description of the first entity, a sequence of the first entity updated text description word granularity semantic feature vector is obtained through an updated text description semantic encoder comprising a word embedding layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311769481.8A CN117610658A (en) | 2023-12-21 | 2023-12-21 | Knowledge graph data dynamic updating method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311769481.8A CN117610658A (en) | 2023-12-21 | 2023-12-21 | Knowledge graph data dynamic updating method and system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117610658A true CN117610658A (en) | 2024-02-27 |
Family
ID=89956291
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311769481.8A Pending CN117610658A (en) | 2023-12-21 | 2023-12-21 | Knowledge graph data dynamic updating method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117610658A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117744785A (en) * | 2024-02-19 | 2024-03-22 | 北京博阳世通信息技术有限公司 | Space-time knowledge graph intelligent construction method and system based on network acquisition data |
-
2023
- 2023-12-21 CN CN202311769481.8A patent/CN117610658A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117744785A (en) * | 2024-02-19 | 2024-03-22 | 北京博阳世通信息技术有限公司 | Space-time knowledge graph intelligent construction method and system based on network acquisition data |
| CN117744785B (en) * | 2024-02-19 | 2024-09-03 | 北京博阳世通信息技术有限公司 | Space-time knowledge graph intelligent construction method and system based on network acquisition data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114201581B (en) | Long text retrieval model based on contrast learning | |
| WO2023065617A1 (en) | Cross-modal retrieval system and method based on pre-training model and recall and ranking | |
| CN109815336B (en) | Text aggregation method and system | |
| CN112749326B (en) | Information processing method, information processing device, computer equipment and storage medium | |
| CN112925904B (en) | Lightweight text classification method based on Tucker decomposition | |
| CN114329109B (en) | Multimodal retrieval method and system based on weakly supervised Hash learning | |
| CN117610658A (en) | Knowledge graph data dynamic updating method and system based on artificial intelligence | |
| CN111723295B (en) | Content distribution method, device and storage medium | |
| CN115294427A (en) | Stylized image description generation method based on transfer learning | |
| CN112597296A (en) | Abstract generation method based on plan mechanism and knowledge graph guidance | |
| CN113821635A (en) | Text abstract generation method and system for financial field | |
| CN117812381B (en) | Video content making method based on artificial intelligence | |
| CN116204694A (en) | Multi-mode retrieval method based on deep learning and hash algorithm | |
| CN107832307B (en) | Chinese word segmentation method based on undirected graph and single-layer neural network | |
| CN117744785B (en) | Space-time knowledge graph intelligent construction method and system based on network acquisition data | |
| CN117407532A (en) | Method for enhancing data by using large model and collaborative training | |
| EP4285281A1 (en) | Annotation-efficient image anomaly detection | |
| CN115064154A (en) | Method and device for generating mixed language voice recognition model | |
| CN117635275B (en) | Intelligent electronic commerce operation commodity management platform and method based on big data | |
| CN112988982B (en) | Autonomous learning method and system for computer comparison space | |
| CN115392254A (en) | Interpretable cognitive prediction and discrimination method and system based on target task | |
| CN111930892A (en) | Scientific and technological text classification method based on improved mutual information function | |
| CN117609527A (en) | Cross-modal data retrieval optimization method based on vector database | |
| CN109657691B (en) | Image semantic annotation method based on energy model | |
| WO2023168818A1 (en) | Method and apparatus for determining similarity between video and text, electronic device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |