Knowledge graph link prediction method and device, equipment, medium and program product

Abstract

The invention discloses a knowledge graph link prediction method, device and equipment, a medium and a program product, and the method comprises the steps: generating a candidate tail entity set based on a head entity element and a relation element of a to-be-queried triple in a knowledge graph, the candidate tail entity set comprising a plurality of candidate tail entity elements, performing semantic analysis on the to-be-queried triad through the language analysis model to obtain semantic information of the to-be-queried triad, screening the candidate tail entity set based on the semantic information, generating a to-be-scored tail entity set according to a screening result, and performing comprehensive scoring on each to-be-scored tail entity element in the to-be-scored tail entity set to obtain a scoring result. According to a comprehensive scoring result, target tail entity elements are screened out from a to-be-scored tail entity set, link prediction is carried out on the knowledge graph based on the target tail entity elements and the to-be-queried triad, potential features in the triad are effectively mined, accurate completion of missing triad entities is achieved, and the accuracy of the whole knowledge graph is improved. And the accuracy of knowledge graph link prediction is greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of knowledge graph technology, and in particular to a knowledge graph link prediction method, device, equipment, medium and program product. Background Art

[0002] As an important technology in the field of artificial intelligence, knowledge graph has been widely used in many scenarios such as knowledge question answering, financial investment consulting, and medical assistance in recent years. Knowledge graph link prediction is a core task in knowledge graph research, which aims to complete the missing entities or relationships in the knowledge graph through reasoning technology to improve its completeness and practicality. The existing knowledge graph link prediction methods are mainly based on triples. However, the current knowledge graph link prediction based on triples cannot effectively mine the potential features in the triples, so it cannot accurately complete the missing entities in the triples, resulting in a low accuracy rate of knowledge graph link prediction, which cannot meet the actual needs of knowledge graph link prediction in the application field. Summary of the invention

[0003] The main purpose of the present invention is to provide a knowledge graph link prediction method, device, equipment, medium and program product, aiming to solve the technical problem that the prior art cannot effectively mine the potential features in triples, and therefore cannot accurately complete the missing entities in the triples, resulting in low accuracy of knowledge graph link prediction.

[0004] To achieve the above object, the present invention provides a knowledge graph link prediction method, which comprises the following steps: Generate a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, wherein the candidate tail entity set includes multiple candidate tail entity elements; Performing semantic analysis on the triple to be queried through a language analysis model to obtain semantic information of the triple to be queried; Screening the candidate tail entity set based on the semantic information, and generating a tail entity set to be scored according to the screening result; Comprehensively scoring each tail entity element to be scored in the tail entity set to be scored, and selecting a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result; Link prediction is performed on the knowledge graph based on the target tail entity element and the triple to be queried.

[0005] Optionally, generating a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph includes: The embedding feature representation model is used to quantify the embedding features of the head entity element and the relationship element of the triple to be queried in the knowledge graph, and the head entity embedding feature vector and the relationship embedding feature vector are obtained; Generate a plurality of original tail entity elements based on the head entity embedding feature vector and the relationship embedding feature vector; Scoring the original tail entity element by using the embedded feature representation model to obtain a first scoring result; Filter the original tail entity elements based on the first scoring result to obtain a first tail entity set; Performing semantic correlation analysis on the head entity element and the relationship element through a language analysis model to obtain semantic related information; Generate a second tail entity set according to the semantic related information; Aggregating the first tail entity set and the second tail entity set to obtain an initial tail entity set; Scoring each initial tail entity element in the initial tail entity set to obtain a second scoring result; A plurality of candidate tail entity elements are screened out from the initial tail entity element based on the second scoring result, and a candidate tail entity set is generated based on the plurality of candidate tail entity elements.

[0006] Optionally, scoring each initial tail entity element in the initial tail entity set to obtain a second scoring result includes: Performing embedding feature quantization on each initial tail entity element in the initial tail entity set by using the embedding feature representation model to obtain an initial tail entity embedding feature vector; Based on the initial tail entity embedding feature vector, the head entity embedding feature vector and the relationship embedding feature vector, the relevance between each initial tail entity element and the triple to be queried is evaluated to obtain a relevance scoring result; Performing diversity scoring on each initial tail entity element according to the cosine similarity between the embedded feature vectors of each initial tail entity to obtain a diversity scoring result; Scoring each initial tail entity element in the initial tail entity set based on the relevance scoring result and the diversity scoring result to obtain a second scoring result: in, Indicates the initial tail attribute element The second scoring result is represents the rating importance weight, Used to control the importance weights of relevance scoring results and diversity scoring results. represents the initial tail entity set, Indicates the initial tail attribute element The relevance score results are: Represents the initial tail entity set Initial tail attribute element Diversity scoring results.

[0007] Optionally, the semantic information includes: context information of the head entity element and semantic description information and adversarial description information of the triple to be queried; The performing semantic analysis on the triple to be queried by using a language analysis model to obtain semantic information of the triple to be queried includes: Input the head entity element and the relationship element into the language analysis model for semantic expansion analysis to obtain context information of the head entity element: in, Represents the header entity element context information, represents the language analysis model, a language hint template representing a language analysis model; Performing semantic description analysis on the triple to be queried based on the head entity element and the relationship element to obtain semantic description information of the triple to be queried: in, Represents the semantic description information of the triple to be queried, Represents the head entity element, Represents a relational element; According to the second scoring result of each candidate tail entity element, an adversarial analysis tail entity element that meets a preset adversarial condition is screened out from the candidate tail entity set; Based on the adversarial analysis tail entity element, the head entity element and the relationship element, the description adversarial analysis is performed on the triple to be queried to obtain adversarial description information of the triple to be queried: in, Represents the adversarial description information of the query triple, Represents the adversarial analysis tail entity element.

[0008] Optionally, the tail entity set to be scored includes a third tail entity set, a fourth tail entity set and a fifth tail entity set; screening the candidate tail entity set based on the semantic information and generating the tail entity set to be scored according to the screening result includes: Generate entity type constraints based on the relationship elements, and perform entity type analysis on each candidate tail entity element according to the entity type constraints to obtain entity type analysis results: in, Indicates the result of entity type analysis. Indicates that the relationship element Next, the head entity element The expected tail entity type of Represents the tail entity type that meets the entity type constraint; Performing entity type screening on the candidate tail entity set according to the entity type constraint condition to obtain a third tail entity set; According to the embedded feature vector of the head entity element, a semantic correlation analysis is performed on each candidate tail entity element in the candidate tail entity set to obtain a semantic correlation analysis result between the head entity element and each candidate tail entity element: in, Represents the result of semantic relevance analysis. represents the embedding feature vector of the head entity element, The embedding feature vector representing the candidate tail entity element; Based on the semantic relevance analysis result, the candidate tail entity set is subjected to semantic relevance screening to obtain a fourth tail entity set; Perform context consistency analysis on each candidate tail entity element in the candidate tail entity set according to the semantic information to obtain a context consistency analysis result between each candidate tail entity element and the knowledge graph: in, Represents the result of context consistency analysis, Representing relationship elements and The semantic similarity between relation elements, represents the knowledge graph, represents the indicator function, Used to judge triples Whether it already exists in the knowledge graph middle; Based on the context consistency analysis result, the candidate tail entity set is screened for context consistency to obtain a fifth tail entity set.

[0009] Optionally, the step of comprehensively scoring each tail entity element to be scored in the tail entity set to be scored, and selecting a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result, comprises: Perform weight matching based on the entity type analysis result, the semantic relevance analysis result, and the context consistency analysis result to obtain a weight parameter; According to the weight parameter, each tail entity element to be scored in the tail entity set to be scored is comprehensively scored to obtain a comprehensive scoring result: in, is the entity type weight parameter, is the semantic relevance weight parameter, is the context consistency weight parameter, Indicates the result of entity type analysis. Represents the result of semantic relevance analysis. Represents the result of context consistency analysis, Indicates the comprehensive scoring result; A target tail entity element is screened out from the set of tail entities to be scored according to the comprehensive scoring result.

[0010] In addition, to achieve the above-mentioned purpose, the present invention also proposes a knowledge graph link prediction device, the knowledge graph link prediction device comprising: A tail entity generation module, used to generate a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, wherein the candidate tail entity set includes multiple candidate tail entity elements; A semantic analysis module, used for performing semantic analysis on the triple to be queried through a language analysis model to obtain semantic information of the triple to be queried; A tail entity screening module, used to screen the candidate tail entity set based on the semantic information, and generate a tail entity set to be scored according to the screening result; A comprehensive scoring module, used for comprehensively scoring each tail entity element to be scored in the tail entity set to be scored, and filtering out a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result; A knowledge graph prediction module is used to perform link prediction on the knowledge graph based on the target tail entity element and the triple to be queried.

[0011] In addition, to achieve the above-mentioned purpose, the present application also proposes a knowledge graph link prediction device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the knowledge graph link prediction method described above.

[0012] In addition, to achieve the above-mentioned purpose, the present application also proposes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the knowledge graph link prediction method described above are implemented.

[0013] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, it implements the steps of the knowledge graph link prediction method described above.

[0014] The present invention generates a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, the candidate tail entity set includes multiple candidate tail entity elements, performs semantic analysis on the triple to be queried through a language analysis model, obtains the semantic information of the triple to be queried, screens the candidate tail entity set based on the semantic information, generates a tail entity set to be scored based on the screening result, performs a comprehensive score on each tail entity element to be scored in the tail entity set to be scored, screens out a target tail entity element from the tail entity set to be scored based on the comprehensive scoring result, and performs link prediction on the knowledge graph based on the target tail entity element and the triple to be queried; since the present invention performs semantic analysis on the query triple through the language analysis model, the candidate tail entity set is screened based on the semantic information, and generates a tail entity set to be scored based on the screening result, the tail entity elements to be scored in the tail entity set to be scored are comprehensively scored, the target tail entity element is screened out from the tail entity set to be scored based on the comprehensive scoring result, and the knowledge graph is predicted based on the target tail entity element and the triple to be queried; The semantic analysis of the group is performed to effectively mine the potential features in the triples, and the candidate tail entity set is screened based on the semantic information, so as to effectively avoid the problems of entity type mismatch and context semantic mismatch between the predicted tail entity element and the triple to be queried, thereby improving the accuracy of entity prediction, and performing comprehensive scoring on the screened tail entity set to be scored, and tail entity screening is performed again based on the scoring results to obtain the target tail entity element. By screening the predicted tail entity elements in stages and multiple times, the entity prediction noise is reduced, the purity of the tail entity set is effectively improved, and the accurate completion of the missing triple entities is achieved, thereby greatly improving the accuracy of knowledge graph link prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0016] Figure 1 It is a schematic diagram of the structure of a knowledge graph link prediction device in a hardware operating environment involved in an embodiment of the present invention; Figure 2 This is a schematic diagram of the flow chart of the first embodiment of the knowledge graph link prediction method of the present invention; Figure 3 It is a flowchart of the second embodiment of the knowledge graph link prediction method of the present invention; Figure 4 It is a flowchart of the third embodiment of the knowledge graph link prediction method of the present invention; Figure 5This is a structural block diagram of the first embodiment of the knowledge graph link prediction device of the present invention.

[0017] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0019] Reference Figure 1 , Figure 1 A schematic diagram of the structure of a knowledge graph link prediction device in a hardware operating environment involved in an embodiment of the present invention.

[0020] like Figure 1 As shown, the knowledge graph link prediction device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, WI-FI) interface). The memory 1005 may be a high-speed random access memory (RandomAccess Memory, RAM) or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0021] Those skilled in the art will understand that Figure 1 The structure shown in does not constitute a limitation on the knowledge graph link prediction device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0022] like Figure 1 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a knowledge graph link prediction program.

[0023] exist Figure 1In the knowledge graph link prediction device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the knowledge graph link prediction device of the present invention can be set in the knowledge graph link prediction device, and the knowledge graph link prediction device calls the knowledge graph link prediction program stored in the memory 1005 through the processor 1001, and executes the knowledge graph link prediction method provided by an embodiment of the present invention.

[0024] The embodiment of the present invention provides a knowledge graph link prediction method, referring to Figure 2 , Figure 2 Schematic diagram of the flow chart of the first embodiment of the knowledge graph link prediction method of the present invention.

[0025] In this embodiment, the knowledge graph link prediction method includes the following steps: Step S10: Generate a candidate tail entity set based on the head entity element and relationship element of the triple to be queried in the knowledge graph.

[0026] It should be understood that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or a terminal electronic device capable of realizing the above functions, etc. The following takes the knowledge graph link prediction device (prediction device) as an example to illustrate this embodiment and the following embodiments.

[0027] It should be noted that the candidate tail entity set includes multiple candidate tail entity elements. The query triplet can be a triplet of the missing tail entity element that needs to be completed. For example, the query triplet is ,in" " is the head entity element, " " is a relational element, " is the missing tail attribute element.

[0028] In some embodiments, the prediction device may generate multiple candidate tail entities by means of embedded feature learning, or may generate multiple candidate tail entities based on a pre-trained large language model, and construct a candidate tail entity set based on the candidate tail entities.

[0029] In some embodiments, the prediction device performs a query triplet. , using embedded feature learning to obtain candidate tail entities based on the embedding model; based on the query head entity and relationship Generate candidate tail entities based on the large language model; and obtain a set of candidate tail entities based on the relevance and diversity of the candidate entities .

[0030] Furthermore, in order to improve the diversity of the candidate entity set and thus improve the prediction accuracy, in some embodiments, the above step S10 may include: Step S101: quantify the embedding features of the head entity element and the relationship element of the triple to be queried in the knowledge graph through the embedding feature representation model to obtain the head entity embedding feature vector and the relationship embedding feature vector; Step S102: generating a plurality of original tail entity elements based on the head entity embedding feature vector and the relationship embedding feature vector; Step S103: scoring the original tail entity element by using the embedded feature representation model to obtain a first scoring result; Step S104: Filtering the original tail entity elements based on the first scoring result to obtain a first tail entity set; Step S105: performing semantic correlation analysis on the head entity element and the relationship element through a language analysis model to obtain semantic related information; Step S106: generating a second tail entity set according to the semantic related information; Step S107: Aggregate the first tail entity set and the second tail entity set to obtain an initial tail entity set; Step S108: scoring each initial tail entity element in the initial tail entity set to obtain a second scoring result; Step S109: screening out a plurality of candidate tail entity elements from the initial tail entity element based on the second scoring result, and generating a candidate tail entity set based on the plurality of candidate tail entity elements.

[0031] It should be noted that the embedded feature representation model may be a SAttLE model. The language analysis model may be a Large Language Model (LLM). The first scoring result may be a scoring result corresponding to each original tail entity element generated based on the embedded feature dimension. The second scoring result may be a scoring result of each initial tail entity element in the initial tail entity set. The initial tail entity set may be a set obtained by aggregating the first tail entity set generated based on the embedded feature dimension and the second tail entity set generated based on the semantic feature dimension.

[0032] It should be understood that this embodiment can generate candidate tail entity elements from the embedding feature dimension and the semantic feature dimension respectively: The prediction device can select the embedded feature representation model SAttLE model to learn the embedded feature representation of the triple, generate multiple original tail entity elements, score the original tail entity elements through the embedded feature representation model, and select the top N original tail entity elements to form the first tail entity set. For example, select the top 200 entities with the highest scores to form the first tail entity set , the scoring function refers to the following formula: in, Representation embedding representation model for triples The rating results, Is the original tail attribute element.

[0033] The head entity element and relationship element in the query vector are input into the language analysis model for semantic analysis, and the second tail entity set is generated based on the semantic related information in the semantic analysis result. This set has potential semantic relevance. For example, the head entity element and relationship element in the query vector are = ("Museum A", "Location"), the language analysis model generates related tail entity elements based on the head entity elements and relationship elements, such as "the city where Museum A is located", "the country where Museum A is located", etc. The process of the language analysis model generating the second tail entity set can be expressed as the following formula: in, Represents the second tail entity set, Represents a language analysis model.

[0034] It can be understood that the prediction device merges the first tail entity set generated based on the embedding feature dimension and the second tail entity set generated based on the semantic feature dimension, and removes duplicate tail entities to obtain an initial tail entity set, referring to the following formula: in, Represents the initial tail entity set.

[0035] It should be noted that in order to improve the purity of the tail entity set, the present embodiment may score and screen the initial tail entity set again. For each initial tail entity element in the initial tail entity set, a score is performed using the embedded feature representation model to evaluate the correlation between the initial tail entity element and the triple to be queried. Entities with high embedding model scores usually have closer correlations with the head entity and relationship. Initial tail entity elements with high embedding feature representation model scores are used as candidate tail entity elements.

[0036] Furthermore, in order to evaluate the relevance and diversity of the initial tail entity element, thereby accurately screening out candidate tail entity elements, the above step S108 may include: Step S1081: performing embedding feature quantization on each initial tail entity element in the initial tail entity set by using the embedding feature representation model to obtain an initial tail entity embedding feature vector; Step S1082: evaluating the relevance between each initial tail entity element and the triple to be queried based on the initial tail entity embedding feature vector, the head entity embedding feature vector and the relationship embedding feature vector to obtain a relevance scoring result; Step S1083: performing diversity scoring on each initial tail entity element according to the cosine similarity between the embedding feature vectors of each initial tail entity to obtain a diversity scoring result; Step S1084: scoring each initial tail entity element in the initial tail entity set based on the relevance scoring result and the diversity scoring result to obtain a second scoring result.

[0037] It should be noted that for each initial tail entity element, the score of the embedding feature representation model is Evaluating their relevance to the query triples, entities with high embedding model scores are usually more closely related to the head entity and relations.

[0038] In order to avoid insufficient semantic coverage due to too similar candidate entities, a diversity function is defined , the diversity score is calculated based on the cosine similarity between the embedding feature vectors of each initial tail entity, thereby ensuring that the candidate entities are concentrated on different semantic categories. Define the scoring function Relevance and diversity scores are performed, as shown below: in, Indicates the initial tail attribute element The second scoring result is represents the rating importance weight, Used to control the importance weights of relevance scoring results and diversity scoring results. represents the initial tail entity set, Indicates the initial tail attribute element The relevance score results are: Represents the initial tail entity set Initial tail attribute element Diversity scoring results.

[0039] Predictive equipment can be maximized The final scoring entity is selected from the initial tail entity set. candidate tail entity elements to form a candidate tail entity set that is both highly relevant and diverse, referring to the following formula: in, represents the initial tail entity set, represents the candidate tail entity set, It can be the number of candidate tail attribute elements, for example It can be 50.

[0040] Step S20: performing semantic analysis on the triple to be queried through a language analysis model to obtain semantic information of the triple to be queried.

[0041] It should be noted that the semantic information may be semantically related information of the triple to be queried, for example, the semantic information may include the head entity semantic information, relationship semantic information, context semantic information, triple description semantic information, and triple adversarial description semantic information of the triple to be queried.

[0042] In some embodiments, the prediction device may design a large language model prompt template based on the head entity elements and relationship elements from three aspects: entity enhancement, triple description enhancement, and adversarial description enhancement, and generate semantic information of the triple to be queried based on the large language model prompt template.

[0043] Step S30: screening the candidate tail entity set based on the semantic information, and generating a tail entity set to be scored according to the screening result.

[0044] It should be noted that the tail entity set to be scored can be an entity set consisting of tail entity elements screened from the candidate tail entity set. For example, the candidate tail entity set includes 200 candidate tail entity elements, and 50 candidate tail entity elements that meet the constraints of the semantic information are screened from the candidate tail entity set based on the semantic information, and these 50 candidate tail entity elements are used as the tail entity elements to be scored, thereby constructing the tail entity set to be scored.

[0045] It can be understood that the prediction device can generate semantic constraints based on semantic information, filter the candidate tail entity set based on the constraints, eliminate the tail entities that do not meet the constraints, filter out the candidate tail entity elements that meet the constraints, and generate a set of tail entities to be scored based on the filtered candidate tail entity elements that meet the constraints, wherein the above constraints may include entity semantic constraints, entity type constraints and semantic context matching constraints.

[0046] Step S40: Comprehensively score each tail entity element to be scored in the tail entity set to be scored, and select a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result.

[0047] It should be noted that the comprehensive scoring can be performed by scoring each to-be-scored tail entity element in the to-be-scored tail entity set from multiple dimensions.

[0048] In some embodiments, the prediction device may evaluate each tail entity element to be scored from three dimensions: entity type matching degree, semantic matching degree, and context consistency, and then perform weighted average of the evaluations of the three dimensions to obtain a comprehensive scoring result.

[0049] Step S50: performing link prediction on the knowledge graph based on the target tail entity element and the triple to be queried.

[0050] It should be noted that the prediction device selects the target tail entity element with the highest evaluation / scoring from the tail entity set to be scored as the tail entity prediction result based on the comprehensive scoring result, so as to complete the query triple, and realize link prediction on the knowledge graph based on the completed triple.

[0051] This embodiment generates a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, the candidate tail entity set includes multiple candidate tail entity elements, performs semantic analysis on the triple to be queried through a language analysis model, obtains semantic information of the triple to be queried, screens the candidate tail entity set based on the semantic information, and generates a tail entity set to be scored based on the screening result, performs a comprehensive score on each tail entity element to be scored in the tail entity set to be scored, screens out a target tail entity element from the tail entity set to be scored based on the comprehensive scoring result, and performs link prediction on the knowledge graph based on the target tail entity element and the triple to be queried; since this embodiment uses a language analysis model to perform semantic analysis on the triple to be queried, the candidate tail entity set is filtered based on the semantic information, and the tail entity set to be scored is generated based on the screening result, the tail entity elements to be scored in the tail entity set to be scored are comprehensively scored, and the target tail entity element is filtered out from the tail entity set to be scored based on the comprehensive scoring result, and the knowledge graph is predicted based on the target tail entity element and the triple to be queried; The tuple is semantically analyzed to effectively mine the potential features in the triple, and the candidate tail entity set is screened based on the semantic information, thereby effectively avoiding the problem of entity type mismatch and context semantic mismatch between the predicted tail entity element and the triple to be queried, thereby improving the accuracy of entity prediction, and performing comprehensive scoring on the screened tail entity set to be scored, and the tail entity is screened again based on the scoring results to obtain the target tail entity element. By screening the predicted tail entity elements in stages and multiple times, the entity prediction noise is reduced, the purity of the tail entity set is effectively improved, and the accurate completion of the missing triple entities is achieved, thereby greatly improving the accuracy of the knowledge graph link prediction.

[0052] refer to Figure 3 , Figure 3Schematic diagram of the flow chart of the second embodiment of the knowledge graph link prediction method of the present invention.

[0053] Based on the above first embodiment, in this embodiment, the semantic information includes: context information of the head entity element and semantic description information and adversarial description information of the triple to be queried; In this embodiment, the above step S20 further includes: Step S21: input the head entity element and the relationship element into a language analysis model for semantic expansion analysis to obtain context information of the head entity element.

[0054] It should be noted that in this embodiment, the prediction device can adopt an entity enhancement strategy based on the query triple , semantically expand the head entity element. Specifically, the large language model prompt template of the language analysis model is used to generate the context information of the head entity. For example, the head entity in the query triple = "b city" and relationship = "capital", the header entity enhancement strategy will provide more background information about "city b", such as whether it is the capital of "country B", a major economic and cultural center, etc. The specific representation is as follows: in, Represents the header entity element context information, represents the language analysis model, A language hint template representing a language analysis model.

[0055] Step S22: performing semantic description analysis on the triple to be queried based on the head entity element and the relationship element to obtain semantic description information of the triple to be queried.

[0056] It should be noted that in this embodiment, the prediction device can use triple description enhancement: for a given head entity and relationship , using a large language model prompt template to generate a triple description. Triple description enhancement provides coherent context information, enabling the model to effectively infer candidate tail entities that conform to the triple logic. For example, given a triple ("b city", "capital", "?"), the generated description can be "b city is an important city, it is the capital of a certain country". It is expressed as follows: in, Represents the semantic description information of the triple to be queried, Represents the head entity element, Represents a relationship element.

[0057] Step S23: selecting adversarial analysis tail entity elements that meet preset adversarial conditions from the candidate tail entity set according to the second scoring results of each candidate tail entity element.

[0058] It should be noted that the tail entity element of the adversarial analysis can be a candidate entity with a lower score ranking in the second scoring result. The adversarial description enhances the candidate that is unlikely to be the target entity, prompting the model to identify the rationality of the candidate entity.

[0059] Step S24: performing description adversarial analysis on the triple to be queried based on the adversarial analysis tail entity element, the head entity element and the relationship element to obtain adversarial description information of the triple to be queried.

[0060] It is understandable that the prediction device in this embodiment can use adversarial description enhancement: for a given head entity and relationship , using a large language model prompt template to generate adversarial descriptions of triples. The adversarial description enhances and highlights candidates that are unlikely to be the target entity, prompting the model to discern the rationality of the candidate entities. By analyzing the lower-ranked candidate entities, reasons are generated for these entities not being suitable as tail entities, thereby improving the prediction accuracy. It is expressed as follows: in, Represents the adversarial description information of the query triple, Represents the adversarial analysis tail entity element.

[0061] This embodiment inputs the head entity element and the relationship element into the language analysis model for semantic expansion analysis, obtains the context information of the head entity element, performs semantic description analysis on the triple to be queried based on the head entity element and the relationship element, obtains the semantic description information of the triple to be queried, screens out the adversarial analysis tail entity element that meets the preset adversarial condition from the candidate tail entity set according to the second scoring result of each candidate tail entity element, performs description adversarial analysis on the triple to be queried based on the adversarial analysis tail entity element, the head entity element and the relationship element, and obtains the adversarial description information of the triple to be queried, thereby analyzing the semantic information of the triple to be queried from multiple dimensions, respectively achieving entity semantic enhancement, triple description enhancement and adversarial description enhancement for the triple to be queried, thereby effectively mining the potential semantic features in the triple, thereby improving the prediction accuracy, and effectively avoiding the problem of semantic irrelevance between the predicted entity and the triple.

[0062] refer to Figure 4 , Figure 4Schematic diagram of the flow chart of the third embodiment of the knowledge graph link prediction method of the present invention.

[0063] Based on the above second embodiment, in this embodiment, the tail entity set to be scored includes a third tail entity set, a fourth tail entity set and a fifth tail entity set; In this embodiment, the above step S30 further includes: Step S301: generating entity type constraints based on the relationship elements, and performing entity type analysis on each candidate tail entity element according to the entity type constraints to obtain entity type analysis results.

[0064] It should be noted that this embodiment can be implemented by designing a multi-round question-and-answer chain of reasoning: the chain of reasoning can provide multi-level verification for the knowledge graph link prediction task by simulating the user's step-by-step analysis and elimination thinking process, thereby improving the model's prediction accuracy for candidate entities. In response to the problem of knowledge graph link prediction, a multi-round question-and-answer chain of reasoning is designed to perform display logic reasoning based on contextual prompt learning from three aspects: entity type matching, semantic relevance analysis, and context consistency check, and gradually screen and optimize the set of tail entity candidates. The designed multi-round question-and-answer template is shown in Table 1 below: It can be understood that in this embodiment, by defining the entity type matching function , filter out the tail attribute elements that meet the entity type constraint conditions from the candidate tail attribute elements, refer to the following formula: in, Indicates the result of entity type analysis. Indicates that the relationship element Next, the head entity element The expected tail entity type of Represents the tail entity type that meets the entity type constraint. If Indicates that the candidate tail entity does not meet the entity type constraint.

[0065] Step S302: Filter the candidate tail entity set by entity type according to the entity type constraint to obtain a third tail entity set.

[0066] It should be noted that the third tail entity set includes candidate tail entity elements that meet the entity type constraint condition.

[0067] Step S303: performing semantic relevance analysis on each candidate tail entity element in the candidate tail entity set according to the embedded feature vector of the head entity element, and obtaining a semantic relevance analysis result between the head entity element and each candidate tail entity element.

[0068] It should be noted that, in this embodiment, by defining a semantic matching function Used to filter the semantic relevance between the head entity and the candidate tail entity: in, Represents the result of semantic relevance analysis. represents the embedding feature vector of the head entity element, represents the embedded feature vector of the candidate tail entity element, If the value of is less than the specified threshold, it is considered the tail entity The semantic information of is not sufficient to support it as a reasonable tail entity, so it is removed.

[0069] Step S304: performing semantic relevance screening on the candidate tail entity set based on the semantic relevance analysis result to obtain a fourth tail entity set.

[0070] It should be noted that the fourth tail entity set includes candidate tail entity elements with strong semantic information relevance.

[0071] Step S305: performing context consistency analysis on each candidate tail entity element in the candidate tail entity set according to the semantic information, and obtaining a context consistency analysis result between each candidate tail entity element and the knowledge graph.

[0072] It should be noted that this embodiment defines a context consistency matching function , so as to analyze the contextual consistency between the candidate tail entity element and the knowledge graph, refer to the following formula: in, Represents the result of context consistency analysis, Representing relationship elements and The semantic similarity between relation elements, represents the knowledge graph, represents the indicator function, Used to judge triples Whether it already exists in the knowledge graph If the triple In the knowledge graph If it exists in the knowledge graph, it returns 1, otherwise it returns 0. The context consistency check reflects whether there is supporting evidence for the candidate entity in the knowledge graph, and can further screen out candidate entities that are semantically reasonable and consistent with the context.

[0073] Step S306: performing context consistency screening on the candidate tail entity set based on the context consistency analysis result to obtain a fifth tail entity set.

[0074] It should be noted that the fifth tail entity set contains candidate tail entity elements that are consistent with the knowledge graph context.

[0075] It can be understood that this embodiment obtains tail entity sets (third tail entity set, fourth tail entity set and fifth tail entity set) obtained by filtering the candidate tail entity sets from the entity type constraint dimension, semantic relevance constraint dimension and context consistency constraint dimension respectively, and avoids scoring the candidate tail entity elements that do not meet the constraints of all dimensions by comprehensively scoring each candidate tail entity element in the tail entity sets of the three dimensions, and screens the scores based on the comprehensive scoring results, which effectively improves the scoring efficiency and accuracy. The highest candidate tail entity element is used as the target tail entity element, thereby ensuring that the target tail entity element is optimal in the entity type constraint dimension, the semantic relevance constraint dimension and the context consistency constraint dimension.

[0076] Further, based on the above embodiment, in order to accurately screen out the target tail entity element, in some embodiments, the above step S40 may include: Step S401: performing weight matching based on the entity type analysis result, the semantic relevance analysis result and the context consistency analysis result to obtain a weight parameter; Step S402: Comprehensively score each tail entity element in the tail entity set to be scored according to the weight parameter to obtain a comprehensive scoring result; Step S403: Filter out target tail entity elements from the to-be-scored tail entity set according to the comprehensive scoring result.

[0077] It should be noted that the prediction device can perform comprehensive scores on the tail entity sets (the third tail entity set, the fourth tail entity set, and the fifth tail entity set) obtained by screening the three dimensions based on the analysis results of the three dimensions, and use the candidate tail entity element with the highest comprehensive score as the target tail entity element: in, is the entity type weight parameter, is the semantic relevance weight parameter, is the context consistency weight parameter, Indicates the result of entity type analysis. Represents the result of semantic relevance analysis. Represents the result of context consistency analysis, Indicates the comprehensive scoring result.

[0078] This embodiment generates entity type constraints based on the relationship elements, and performs entity type analysis on each candidate tail entity element according to the entity type constraints to obtain entity type analysis results, performs entity type screening on the candidate tail entity set according to the entity type constraints to obtain a third tail entity set, performs semantic relevance analysis on each candidate tail entity element in the candidate tail entity set according to the embedded feature vector of the head entity element to obtain a semantic relevance analysis result between the head entity element and each candidate tail entity element, performs semantic relevance screening on the candidate tail entity set based on the semantic relevance analysis result to obtain a fourth tail entity set, performs context consistency analysis on each candidate tail entity element in the candidate tail entity set according to the semantic information to obtain a context consistency analysis result between each candidate tail entity element and the knowledge graph, performs context consistency screening on the candidate tail entity set based on the context consistency analysis result to obtain a fifth tail entity set, thereby effectively improving the purity of the candidate tail entity set, and based on multiple screening, through the complementary screening targets at different stages, effectively reducing the scoring complexity and analysis noise, greatly improving the scoring accuracy and scoring quality, and thus effectively improving the prediction accuracy.

[0079] In addition, an embodiment of the present invention also proposes a computer-readable storage medium, on which a knowledge graph link prediction program is stored. When the knowledge graph link prediction program is executed by a processor, the steps of the knowledge graph link prediction method described above are implemented.

[0080] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0081] The above-mentioned computer-readable storage medium may be included in the knowledge graph link prediction device; or it may exist independently without being assembled into the knowledge graph link prediction device.

[0082] In addition, an embodiment of the present invention also proposes a computer program product, including a knowledge graph link prediction program, which, when executed by a processor, implements the steps of the knowledge graph link prediction method described above.

[0083] The specific implementation methods of the computer program product of the present invention are basically the same as the various embodiments of the above-mentioned knowledge graph link prediction method, and will not be repeated here.

[0084] Reference Figure 5 , Figure 5 This is a structural block diagram of the first embodiment of the knowledge graph link prediction device of the present invention.

[0085] like Figure 5 As shown, the knowledge graph link prediction device proposed in the embodiment of the present invention includes: A tail entity generation module 10 is used to generate a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, wherein the candidate tail entity set includes multiple candidate tail entity elements; A semantic analysis module 20, configured to perform semantic analysis on the triple to be queried through a language analysis model to obtain semantic information of the triple to be queried; A tail entity screening module 30, configured to screen the candidate tail entity set based on the semantic information, and generate a tail entity set to be scored according to the screening result; A comprehensive scoring module 40 is used to comprehensively score each tail entity element to be scored in the tail entity set to be scored, and filter out a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result; The knowledge graph prediction module 50 is used to perform link prediction on the knowledge graph based on the target tail entity element and the triple to be queried.

[0086] This embodiment generates a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, the candidate tail entity set includes multiple candidate tail entity elements, performs semantic analysis on the triple to be queried through a language analysis model, obtains semantic information of the triple to be queried, screens the candidate tail entity set based on the semantic information, and generates a tail entity set to be scored based on the screening result, performs a comprehensive score on each tail entity element to be scored in the tail entity set to be scored, screens out a target tail entity element from the tail entity set to be scored based on the comprehensive scoring result, and performs link prediction on the knowledge graph based on the target tail entity element and the triple to be queried; since this embodiment uses a language analysis model to perform semantic analysis on the triple to be queried, the candidate tail entity set is filtered based on the semantic information, and the tail entity set to be scored is generated based on the screening result, the tail entity elements to be scored in the tail entity set to be scored are comprehensively scored, and the target tail entity element is filtered out from the tail entity set to be scored based on the comprehensive scoring result, and the knowledge graph is predicted based on the target tail entity element and the triple to be queried; The tuple is semantically analyzed to effectively mine the potential features in the triple, and the candidate tail entity set is screened based on the semantic information, thereby effectively avoiding the problem of entity type mismatch and context semantic mismatch between the predicted tail entity element and the triple to be queried, thereby improving the accuracy of entity prediction, and performing comprehensive scoring on the screened tail entity set to be scored, and the tail entity is screened again based on the scoring results to obtain the target tail entity element. By screening the predicted tail entity elements in stages and multiple times, the entity prediction noise is reduced, the purity of the tail entity set is effectively improved, and the accurate completion of the missing triple entities is achieved, thereby greatly improving the accuracy of the knowledge graph link prediction.

[0087] The knowledge graph link prediction device provided by this application adopts the knowledge graph link prediction method in the above embodiment, which can solve the technical problem of knowledge graph link prediction. Compared with the prior art, the beneficial effects of the knowledge graph link prediction device provided by this application are the same as the beneficial effects of the knowledge graph link prediction method provided by the above embodiment, and the other technical features in the knowledge graph link prediction device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0088] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0089] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0090] In addition, for technical details not fully described in this embodiment, please refer to the knowledge graph link prediction method provided in any embodiment of the present invention, and will not be repeated here.

[0091] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0092] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0093] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0094] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A knowledge graph link prediction method, characterized in that: The knowledge graph link prediction method comprises: Generate a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, wherein the candidate tail entity set includes multiple candidate tail entity elements; Performing semantic analysis on the triple to be queried through a language analysis model to obtain semantic information of the triple to be queried; Screening the candidate tail entity set based on the semantic information, and generating a tail entity set to be scored according to the screening result; Comprehensively scoring each tail entity element to be scored in the tail entity set to be scored, and selecting a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result; Link prediction is performed on the knowledge graph based on the target tail entity element and the triple to be queried.

2. The knowledge graph link prediction method according to claim 1, characterized in that: The step of generating a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph includes: The embedding feature representation model is used to quantify the embedding features of the head entity element and the relationship element of the triple to be queried in the knowledge graph, and the head entity embedding feature vector and the relationship embedding feature vector are obtained; Generate a plurality of original tail entity elements based on the head entity embedding feature vector and the relationship embedding feature vector; Scoring the original tail entity element by using the embedded feature representation model to obtain a first scoring result; Filter the original tail entity elements based on the first scoring result to obtain a first tail entity set; Performing semantic correlation analysis on the head entity element and the relationship element through a language analysis model to obtain semantic related information; Generate a second tail entity set according to the semantic related information; Aggregating the first tail entity set and the second tail entity set to obtain an initial tail entity set; Scoring each initial tail entity element in the initial tail entity set to obtain a second scoring result; A plurality of candidate tail entity elements are screened out from the initial tail entity element based on the second scoring result, and a candidate tail entity set is generated based on the plurality of candidate tail entity elements.

3. The knowledge graph link prediction method according to claim 2, characterized in that: Scoring each initial tail entity element in the initial tail entity set to obtain a second scoring result includes: Performing embedding feature quantization on each initial tail entity element in the initial tail entity set by using the embedding feature representation model to obtain an initial tail entity embedding feature vector; Based on the initial tail entity embedding feature vector, the head entity embedding feature vector and the relationship embedding feature vector, the relevance between each initial tail entity element and the triple to be queried is evaluated to obtain a relevance scoring result; Performing diversity scoring on each initial tail entity element according to the cosine similarity between the embedded feature vectors of each initial tail entity to obtain a diversity scoring result; Scoring each initial tail entity element in the initial tail entity set based on the relevance scoring result and the diversity scoring result to obtain a second scoring result: in, Indicates the initial tail attribute element The second scoring result is represents the rating importance weight, Used to control the importance weights of relevance scoring results and diversity scoring results. represents the initial tail entity set, Indicates the initial tail attribute element The relevance score results are: Represents the initial tail entity set Initial tail attribute element Diversity scoring results.

4. The knowledge graph link prediction method according to any one of claims 1 to 3, characterized in that: The semantic information includes: context information of the head entity element and semantic description information and adversarial description information of the triple to be queried; The performing semantic analysis on the triple to be queried by using a language analysis model to obtain semantic information of the triple to be queried includes: Input the head entity element and the relationship element into the language analysis model for semantic expansion analysis to obtain context information of the head entity element: in, Represents the header entity element Context information, represents the language analysis model, a language hint template representing a language analysis model; Performing semantic description analysis on the triple to be queried based on the head entity element and the relationship element to obtain semantic description information of the triple to be queried: in, Represents the semantic description information of the triple to be queried, Represents the head entity element, Represents a relational element; According to the second scoring result of each candidate tail entity element, an adversarial analysis tail entity element that meets a preset adversarial condition is screened out from the candidate tail entity set; Based on the adversarial analysis tail entity element, the head entity element and the relationship element, the description adversarial analysis is performed on the triple to be queried to obtain adversarial description information of the triple to be queried: in, Represents the adversarial description information of the query triple, Represents the adversarial analysis tail entity element.

5. The knowledge graph link prediction method according to claim 4, characterized in that: The tail entity set to be scored includes a third tail entity set, a fourth tail entity set and a fifth tail entity set; the screening of the candidate tail entity set based on the semantic information and generating the tail entity set to be scored according to the screening result include: Generate entity type constraints based on the relationship elements, and perform entity type analysis on each candidate tail entity element according to the entity type constraints to obtain entity type analysis results: in, Indicates the result of entity type analysis. Indicates that the relationship element Next, the head entity element The expected tail entity type of Represents the tail entity type that meets the entity type constraint; Performing entity type screening on the candidate tail entity set according to the entity type constraint condition to obtain a third tail entity set; According to the embedded feature vector of the head entity element, a semantic correlation analysis is performed on each candidate tail entity element in the candidate tail entity set to obtain a semantic correlation analysis result between the head entity element and each candidate tail entity element: in, Represents the result of semantic relevance analysis. represents the embedding feature vector of the head entity element, The embedding feature vector representing the candidate tail entity element; Based on the semantic relevance analysis result, the candidate tail entity set is subjected to semantic relevance screening to obtain a fourth tail entity set; Perform context consistency analysis on each candidate tail entity element in the candidate tail entity set according to the semantic information to obtain a context consistency analysis result between each candidate tail entity element and the knowledge graph: in, Represents the result of context consistency analysis, Representing relationship elements and The semantic similarity between relation elements, represents the knowledge graph, represents the indicator function, Used to judge triples Whether it already exists in the knowledge graph middle; Based on the context consistency analysis result, the candidate tail entity set is screened for context consistency to obtain a fifth tail entity set.

6. The knowledge graph link prediction method according to claim 5, characterized in that: The step of comprehensively scoring each tail entity element to be scored in the tail entity set to be scored, and selecting a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result, comprises: Perform weight matching based on the entity type analysis result, the semantic relevance analysis result, and the context consistency analysis result to obtain a weight parameter; According to the weight parameter, each tail entity element to be scored in the tail entity set to be scored is comprehensively scored to obtain a comprehensive scoring result: in, is the entity type weight parameter, is the semantic relevance weight parameter, is the context consistency weight parameter, Indicates the result of entity type analysis. Represents the result of semantic relevance analysis. Represents the result of context consistency analysis, Indicates the comprehensive scoring result; A target tail entity element is screened out from the set of tail entities to be scored according to the comprehensive scoring result.

7. A knowledge graph link prediction device, characterized in that: The knowledge graph link prediction device comprises: A tail entity generation module, used to generate a candidate tail entity set based on the head entity element and the relationship element of the triple to be queried in the knowledge graph, wherein the candidate tail entity set includes multiple candidate tail entity elements; A semantic analysis module, used for performing semantic analysis on the triple to be queried through a language analysis model to obtain semantic information of the triple to be queried; A tail entity screening module, used to screen the candidate tail entity set based on the semantic information, and generate a tail entity set to be scored according to the screening result; A comprehensive scoring module, used for comprehensively scoring each tail entity element to be scored in the tail entity set to be scored, and filtering out a target tail entity element from the tail entity set to be scored according to the comprehensive scoring result; A knowledge graph prediction module is used to perform link prediction on the knowledge graph based on the target tail entity element and the triple to be queried.

8. A knowledge graph link prediction device, characterized in that: The knowledge graph link prediction device includes: a memory, a processor, and a knowledge graph link prediction program stored in the memory and executable on the processor, wherein the knowledge graph link prediction program is configured to implement the knowledge graph link prediction method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a knowledge graph link prediction program, and when the knowledge graph link prediction program is executed by the processor, the knowledge graph link prediction method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product includes a knowledge graph link prediction program, which, when executed by a processor, implements the steps of the knowledge graph link prediction method according to any one of claims 1 to 6.

Images