CN118036732A - Social event pattern relation completion method and system based on critical countermeasure learning - Google Patents
Social event pattern relation completion method and system based on critical countermeasure learning Download PDFInfo
- Publication number
- CN118036732A CN118036732A CN202410430489.XA CN202410430489A CN118036732A CN 118036732 A CN118036732 A CN 118036732A CN 202410430489 A CN202410430489 A CN 202410430489A CN 118036732 A CN118036732 A CN 118036732A
- Authority
- CN
- China
- Prior art keywords
- critical
- relationship
- knowledge graph
- node
- graph
- 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
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000006870 function Effects 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims description 37
- 238000013528 artificial neural network Methods 0.000 claims description 19
- 239000013598 vector Substances 0.000 claims description 16
- 238000013507 mapping Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 5
- 230000004931 aggregating effect Effects 0.000 claims 1
- 238000012512 characterization method Methods 0.000 abstract description 4
- 238000012549 training Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the technical field of knowledge maps, and particularly discloses a social event map relation completion method and system based on critical countermeasure learning, wherein the method comprises the following steps: acquiring a knowledge graph of historical social event information; any node pair in the knowledge graphInputting the information into a trained knowledge-graph relation prediction model, and outputting nodesIs in between withScore of a relationship, determining whether there is a relationship between two nodes based on the scoreRelationship, carrying out relationship completion in the knowledge graph; the loss function of the knowledge graph relation prediction model comprises critical countermeasures. According to the method, critical countermeasures based on the triplet feature distance sorting are introduced into the loss function, so that the certainty of the fuzzy triplet approximation relationship is higher, the embedding characterization capability of the triplet is improved, and the prediction performance of the knowledge graph relationship prediction model is improved.
Description
Technical Field
The invention relates to the technical field of knowledge graphs, in particular to a social event graph relationship completion method and system based on critical countermeasure learning.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
In the process of resolving social contradictions, social event knowledge graphs are usually required to be constructed for analyzing risks, but the knowledge graphs are often incomplete, and a plurality of missing relations or nodes exist.
The map link prediction, also called map completion, is an automatic reasoning technology for predicting missing links in a knowledge map. The existing map link prediction method has four directions, namely: vector conversion-based models, semantic matching-based models, convolutional neural network-based models, and graph neural network-based models. Recently, many approaches have focused on knowledge graph embedding, mapping entities and relationships to low-dimensional vector space through sub-graph aggregation, while preserving structural and semantic information as much as possible. These most advanced methods are mainly based on graph embedding predicting the link scores between source and target nodes, but ignore the meaning of triplet embedding in different interactive subgraphs. However, the manner of interaction between triplets is different, and the link prediction model may lack interactivity, i.e., noise and uncertainty exists in the embedded representation of the knowledge graph. Since link prediction is implemented based on graph embedding, these embedded noises seriously affect the performance of link prediction.
Disclosure of Invention
In order to solve the problems, the invention provides a social event pattern relation completion method and a system based on critical countermeasure learning, which are used for increasing critical countermeasure loss in a loss function when training a knowledge pattern relation prediction model, so that the model can improve the embedding representation capability of triples in the knowledge pattern while finding similar or different meanings of the triples.
In some embodiments, the following technical scheme is adopted:
A social event pattern relation completion method based on critical countermeasure learning comprises the following steps:
acquiring knowledge graph of historical social event information Wherein/>For node collection,/>For edge set,/>Is a collection of relationship types;
Mapping knowledge graph Pair of arbitrary nodes/>Inputting the information into a trained knowledge graph relation prediction model, and outputting nodes/>Exist/>Score of relationship, based on the score, determining whether there is/>, between two nodesRelationship, carrying out relationship completion in the knowledge graph;
The loss function of the knowledge graph relation prediction model comprises critical countermeasures loss, and the calculation process of the critical countermeasures loss is as follows: calculating cosine similarity of any two triples in the triplet set, and constructing a cosine similarity matrix; the values in each row and each column of the cosine similarity matrix are arranged in a descending order, and then backward difference is carried out on each row to obtain a difference matrix; and determining the position of the maximum value in each row and each column of the differential matrix, and calculating critical countermeasure loss by combining the element values in the differential matrix.
Wherein, the critical countermeasures loss are specifically:
Wherein, And/>Is super-parameter,/>For the number of triples in the knowledge graph,/>For the position of the maximum value in each row and column of the differential matrix,/>The elements in the matrix obtained by descending order of the values in each row and column in the cosine similarity matrix are obtained.
In other embodiments, the following technical solutions are adopted:
A social event graph relationship completion system based on critical countermeasure learning, comprising:
the knowledge graph data acquisition module is used for acquiring knowledge graphs of the historical social event information Wherein/>For node collection,/>For edge set,/>Is a collection of relationship types;
the knowledge graph relation complementing module is used for complementing the knowledge graph Pair of arbitrary nodes/>Inputting the information into a trained knowledge graph relation prediction model, and outputting nodes/>Exist/>Score of relationship, based on the score, determining whether there is/>, between two nodesRelationship, carrying out relationship completion in the knowledge graph;
The loss function of the knowledge graph relation prediction model comprises critical countermeasures loss, and the calculation process of the critical countermeasures loss is as follows: calculating cosine similarity of any two triples in the triplet set, and constructing a cosine similarity matrix; the values in each row and each column of the cosine similarity matrix are arranged in a descending order, and then backward difference is carried out on each row to obtain a difference matrix; and determining the position of the maximum value in each row and each column of the differential matrix, and calculating critical countermeasure loss by combining the element values in the differential matrix.
In other embodiments, the following technical solutions are adopted:
A terminal device comprising a processor and a memory, the processor for implementing instructions; the memory is used for storing a plurality of instructions which are suitable for being loaded by the processor and executing the social event pattern relation complementing method based on critical countermeasure learning.
In other embodiments, the following technical solutions are adopted:
A computer readable storage medium having stored therein a plurality of instructions adapted to be loaded by a processor of a terminal device and to perform the above-described critical countermeasure learning based social event pattern relationship completion method.
Compared with the prior art, the invention has the beneficial effects that:
(1) According to the method, critical countermeasures based on the sorting of the feature distances of the triples are introduced into the loss function, the semantic similarity of the triples is sorted, the position with the strongest gradient change in theory is the position with the strongest relative semantic uncertainty of the triples, the position with the strongest certainty is the forefront position on the curve, namely the position with the strongest certainty, and the curve is steeper through countermeasures, namely the certainty of the approximate relation of the blurred triples is stronger, so that the embedding characterization capacity of the triples is improved, and the prediction performance of the knowledge graph relation prediction model is improved.
(2) According to the method, the cross entropy loss is obtained by obtaining the expectation of the cross entropy between the prediction results of all candidate triples in the knowledge graph and the real sample; the cross entropy loss is a reference loss function based on the original knowledge graph structure, and is an important component of overall loss; the method combines the cross loss and the critical countermeasure loss to obtain the overall loss function of the knowledge graph relation prediction model, and can solve the problem of inaccurate prediction results caused by noise and uncertainty existing in the embedding characterization of the knowledge graph.
Additional features and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
FIG. 1 is a schematic diagram of a social event pattern relationship completion method based on critical countermeasure learning in an embodiment of the invention;
FIG. 2 (a) is a sub-graph randomly extracted from a knowledge-graph in an embodiment of the present invention;
FIG. 2 (B) is a node associated with node B in the extracted subgraph;
Fig. 2 (c) is a node associated with node pair (a, B) in the extracted subgraph.
Detailed Description
It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present application. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.
Example 1
In one or more embodiments, a social event pattern relationship completion method based on critical countermeasure learning is disclosed, and in combination with fig. 1, the method specifically includes the following steps:
(1) Acquiring knowledge graph of historical social event information Wherein/>For node collection,/>For edge set,/>Is a collection of relationship types;
it should be noted that, knowledge graph of social event information A knowledge graph for representing the association relationship between the event related nodes, for example: /(I)The system consists of event type, place, time, disposal personnel, war technique and tactics. Relation set/>The method is composed of relations such as event-event identity, event-event homology, treatment-event and the like. /(I)Indicating whether a specified relationship exists between any two nodes, wherein the existence is 1, otherwise, the existence is 0.
Whereby triples can be constructedWherein/>For node set/>Is one of any two nodes,/>Representing a relationship between two nodes; an example triplet is given below: (Zhang three, handle-event, event A).
(2) Mapping knowledge graphPair of arbitrary nodes/>Inputting the information into a trained knowledge graph relation prediction model, and outputting nodes/>Exist/>Score of a relationship, determining whether there is a relationship between two nodes based on the scoreThe relationship, i.e. when the score is greater than a certain threshold, it is assumed that there is/>, between two nodesThe relationship, the knowledge graph needs to be complemented accordingly; the specific threshold value can be selected by those skilled in the art according to actual needs.
In this embodiment, the knowledge graph relationship predicts the loss function of the modelIs cross entropy loss/>, based on the presence or absence of triplesAnd critical countermeasure loss/>, based on triplet feature distance orderingAnd (2) sum:
;
By using And training the knowledge graph relation prediction model by using the loss function until convergence to obtain a trained knowledge graph relation prediction model.
In this embodiment, knowledge graph based on social event informationNode set/>Middle/>Personal nodeThe text information of (a) is converted into text feature vectors to obtain the (a) >Initial eigenvector/>, of individual node,/>; Using node pairs/>Representing the relationship between two nodes/>。
Node is connected withThe initial feature vector of (1) is input to a graph neural network with an attention mechanism to obtain a node/>Embedded features/>; Node/>The initial feature vector of (1) is input to a graph neural network with an attention mechanism to obtain a node/>Embedded features/>; Node/>And node/>The initial feature vector of (1) is input to a graph neural network with an attention mechanism to obtain node pairs/>Embedded features/>;
Will embed features、/>And/>Input to a trained transducer-based neural network, get representation triplet/>Fusion characteristics/>。
As a specific example, a subgraph is randomly extracted from the knowledge graph, and fig. 2 (a) shows an example of a subgraph, where the subgraph includes A, B, C, D, E nodes, and initial feature vectors of each node are respectively X A、XB、XC、XD and X E; as shown in fig. 2 (B) and 2 (C), in this subgraph, the nodes associated with node B are node D and node E, and the nodes associated with node pair (a, B) are node C, node D and node E.
In this embodiment, GCN (Graph Convolutional Network) neural networks are adopted, the input of the GCN neural network is the initial feature vector of the node, and the output is the embedded feature of the node. The GCN neural network will use the node information associated with the node to determine the embedded characteristics of the node.
With reference to fig. 1, an initial feature vector X A of the node a is input into the GCN neural network to obtain an embedded feature of the node a; the GCN neural network obtains the embedded characteristics of the node A by utilizing the related information of the node C and the node D related to the node A; Similarly, the initial feature vector X B of the node B is input into the GCN neural network to obtain the embedded feature/>, of the node B; The initial feature vectors of the node A and the node B are simultaneously input into the GCN neural network, so that the embedded feature/> -of the node pair (A, B) can be obtained。
Will embed features、/>And/>Input to a trained transducer-based neural network (e.g., BERT (Bidirectional Encoder Representations from Transformers, bi-directional encoder representation from transducer) may be chosen to obtain a representation triplet/>Fusion characteristics/>。
In the present embodiment, forArbitrary relation/>,/>For its learnable parameter matrix, will/>Mapping to/>Obtain mapping score/>By/>Function pair/>Normalized to give/>:
;
;
Sampling method is adopted to obtain knowledge graphObtaining the triplet set/>,/>For Batchsize, the number of triples in the triplet set of the knowledge graph, the node relation without edges is defined as Nan,/>Use/>Judging candidate triples/>Whether the knowledge graph exists truly or not, and calculating/>Cross entropy loss/>:
;
Wherein,。
Further, calculateIs/are of any tripletAnd/>Cosine similarity/>Obtaining a matrixTo/>The middle row is used as a unit to form a plurality of columns, and the values in each row and each column are arranged in descending order to obtain:
;
;
;
For a pair ofBackward difference is carried out on each row in the array to obtain/>:
;
;
SelectingFirst/>The row number column, the position of the greatest value in the number column/>, is obtained(E.g., 200 total values in the series, with the 100 th value being the largest, then/>) Calculate critical countermeasures loss/>:
;
Wherein,And/>Is a super parameter and can be set according to the requirement; /(I)For the number of triples in the knowledge graph,/>For the position of the maximum value in each row and column of the differential matrix,/>The elements in the matrix obtained by descending order of the values in each row and column in the cosine similarity matrix are obtained.
When the application model is used for reasoning, any node pair of the knowledge graph is inputThe relation/>, exists between two output nodesWherein/>. When the score is greater than a certain threshold, it is determined that there is/>, between the two nodesThe relationship needs to complement the knowledge graph accordingly.
The method can improve the embedding characterization capability of the triples in the knowledge graph, and further improve the performance of the knowledge graph relation prediction model.
Example two
In one or more embodiments, a social event graph relationship completion system based on critical countermeasure learning is disclosed, comprising:
the knowledge graph data acquisition module is used for acquiring knowledge graphs of the historical social event information Wherein/>For node collection,/>For edge set,/>Is a collection of relationship types;
the knowledge graph relation complementing module is used for complementing the knowledge graph Pair of arbitrary nodes/>Inputting the information into a trained knowledge graph relation prediction model, and outputting nodes/>Exist/>Score of relationship, based on the score, determining whether there is/>, between two nodesRelationship, carrying out relationship completion in the knowledge graph;
The loss function of the knowledge graph relation prediction model comprises critical countermeasures loss, and the calculation process of the critical countermeasures loss is as follows: calculating cosine similarity of any two triples in the triplet set, and constructing a cosine similarity matrix; the values in each row and each column of the cosine similarity matrix are arranged in a descending order, and then backward difference is carried out on each row to obtain a difference matrix; and determining the position of the maximum value in each row and each column of the differential matrix, and calculating critical countermeasure loss by combining the element values in the differential matrix.
The specific implementation of each module is the same as that in the first embodiment, and will not be described in detail.
Example III
In one or more embodiments, a terminal device is disclosed, including a server including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the critical countermeasure learning-based social event pattern relationship completion method in embodiment one when executing the program. For brevity, the description is omitted here.
It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general purpose processors, digital signal processors DSP, application specific integrated circuits ASIC, off-the-shelf programmable gate array FPGA or other programmable logic device, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory may include read only memory and random access memory and provide instructions and data to the processor, and a portion of the memory may also include non-volatile random access memory. For example, the memory may also store information of the device type.
In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software.
Example IV
In one or more embodiments, a computer-readable storage medium is disclosed, in which a plurality of instructions are stored, the instructions being adapted to be loaded by a processor of a terminal device and to perform the critical countermeasure learning-based social event graph relationship completion method described in embodiment one.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (10)
1. A social event pattern relation completion method based on critical countermeasure learning is characterized by comprising the following steps:
acquiring knowledge graph of historical social event information Wherein/>For node collection,/>For edge set,/>Is a collection of relationship types;
Mapping knowledge graph Pair of arbitrary nodes/>Inputting the information into a trained knowledge graph relation prediction model, and outputting nodes/>Exist/>Score of relationship, based on the score, determining whether there is/>, between two nodesRelationship, carrying out relationship completion in the knowledge graph;
The loss function of the knowledge graph relation prediction model comprises critical countermeasures loss, and the calculation process of the critical countermeasures loss is as follows: calculating cosine similarity of any two triples in the triplet set, and constructing a cosine similarity matrix; the values in each row and each column of the cosine similarity matrix are arranged in a descending order, and then backward difference is carried out on each row to obtain a difference matrix; and determining the position of the maximum value in each row and each column of the differential matrix, and calculating critical countermeasure loss by combining the element values in the differential matrix.
2. The social event pattern relation completion method based on critical countermeasure learning as claimed in claim 1, wherein the critical countermeasure loss is specifically:
Wherein, And/>Is super-parameter,/>For the number of triples in the knowledge graph,/>For the position of the maximum value in each row and column of the differential matrix,/>The elements in the matrix obtained by descending order of the values in each row and column in the cosine similarity matrix are obtained.
3. The social event graph relationship completion method based on critical countermeasure learning as claimed in claim 1, wherein the loss function of the knowledge graph relationship prediction model is: the sum of cross entropy loss and critical contrast loss.
4. The social event graph relationship completion method based on critical countermeasure learning as claimed in claim 3, wherein the cross entropy loss calculation method is as follows:
aggregating nodes Middle/>Individual node/>The text information of (a) is converted into text feature vectors to obtain nodes/>Initial feature vector/>;
Using node pairsRepresenting the relationship between two nodes/>;
Nodes are respectively obtained by adopting graph neural network with attention mechanismNode/>And node pair/>Is embedded with features of (a);
Based on the embedded features, a transducer-based neural network is adopted to obtain triplets Fusion characteristics/>;
For a pair ofArbitrary relation/>,/>Fusion features/>, for its learnable parameter matrixMapping to/>Obtaining a mapping scorePair/>Normalized to give/>Based on/>The cross entropy loss is calculated.
5. The method for completing social event pattern relationships based on critical countermeasure learning according to claim 4, wherein the mapping scoreThe method comprises the following steps:
。
6. The social event graph relationship completion method based on critical countermeasure learning as claimed in claim 4, wherein the cross entropy loss is specifically:
;
Wherein, To from knowledge graph/>And (3) the triplet set obtained in (I)For the number of triples in the knowledge graph,/>For judging candidate triples/>Whether or not the information exists in the knowledge graph truly, if/>=1, Then/>Exist/>=0, Then/>Is not present.
7. The social event graph relationship completion method based on critical countermeasure learning as claimed in claim 4, wherein a triplet is obtainedThe specific process is as follows:
Node is connected with The initial feature vector of (1) is input to a graph neural network with an attention mechanism to obtain a node/>Embedded features of (a);
Node is connected withThe initial feature vector of (1) is input to a graph neural network with an attention mechanism to obtain a node/>Embedded features of (a);
Node is connected withAnd node/>The initial feature vector of (1) is input to a graphic neural network with an attention mechanism to obtain node pairsEmbedded features/>;
Will embed features、/>And/>Input to a trained transducer-based neural network, get representation triplet/>Fusion characteristics/>。
8. A social event graph relationship completion system based on critical countermeasure learning, comprising:
the knowledge graph data acquisition module is used for acquiring knowledge graphs of the historical social event information Wherein/>For node collection,/>For edge set,/>Is a collection of relationship types;
the knowledge graph relation complementing module is used for complementing the knowledge graph Pair of arbitrary nodes/>Inputting the information into a trained knowledge graph relation prediction model, and outputting nodes/>Exist/>Score of relationship, based on the score, determining whether there is/>, between two nodesRelationship, carrying out relationship completion in the knowledge graph;
The loss function of the knowledge graph relation prediction model comprises critical countermeasures loss, and the calculation process of the critical countermeasures loss is as follows: calculating cosine similarity of any two triples in the triplet set, and constructing a cosine similarity matrix; the values in each row and each column of the cosine similarity matrix are arranged in a descending order, and then backward difference is carried out on each row to obtain a difference matrix; and determining the position of the maximum value in each row and each column of the differential matrix, and calculating critical countermeasure loss by combining the element values in the differential matrix.
9. A terminal device comprising a processor and a memory, the processor for implementing instructions; the memory for storing a plurality of instructions adapted to be loaded by the processor and to perform the critical countermeasure learning based social event graph relationship completion method of any of claims 1-7.
10. A computer readable storage medium having stored therein a plurality of instructions adapted to be loaded by a processor of a terminal device and to perform the critical countermeasure learning based social event graph relationship completion method of any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410430489.XA CN118036732A (en) | 2024-04-11 | 2024-04-11 | Social event pattern relation completion method and system based on critical countermeasure learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410430489.XA CN118036732A (en) | 2024-04-11 | 2024-04-11 | Social event pattern relation completion method and system based on critical countermeasure learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN118036732A true CN118036732A (en) | 2024-05-14 |
Family
ID=90989656
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410430489.XA Pending CN118036732A (en) | 2024-04-11 | 2024-04-11 | Social event pattern relation completion method and system based on critical countermeasure learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN118036732A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110147450A (en) * | 2019-05-06 | 2019-08-20 | 北京科技大学 | A kind of the knowledge complementing method and device of knowledge mapping |
CN110472067A (en) * | 2019-08-19 | 2019-11-19 | 腾讯科技(深圳)有限公司 | Knowledge mapping indicates learning method, device, computer equipment and storage medium |
KR102134405B1 (en) * | 2019-06-27 | 2020-07-15 | 중앙대학교 산학협력단 | System and Method for Improving Low Light Level Image Using Generative Adversarial Network |
CN116188543A (en) * | 2022-12-27 | 2023-05-30 | 中国人民解放军61363部队 | Point cloud registration method and system based on deep learning unsupervised |
CN116312484A (en) * | 2023-05-18 | 2023-06-23 | 南京邮电大学 | Cross-language domain invariant acoustic feature extraction method and system |
CN116340540A (en) * | 2023-03-24 | 2023-06-27 | 南京理工大学 | Method for generating network security emergency response knowledge graph based on text |
CN117668256A (en) * | 2023-12-12 | 2024-03-08 | 大连理工大学 | Knowledge graph completion method based on global and local contrast learning |
-
2024
- 2024-04-11 CN CN202410430489.XA patent/CN118036732A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110147450A (en) * | 2019-05-06 | 2019-08-20 | 北京科技大学 | A kind of the knowledge complementing method and device of knowledge mapping |
KR102134405B1 (en) * | 2019-06-27 | 2020-07-15 | 중앙대학교 산학협력단 | System and Method for Improving Low Light Level Image Using Generative Adversarial Network |
CN110472067A (en) * | 2019-08-19 | 2019-11-19 | 腾讯科技(深圳)有限公司 | Knowledge mapping indicates learning method, device, computer equipment and storage medium |
CN116188543A (en) * | 2022-12-27 | 2023-05-30 | 中国人民解放军61363部队 | Point cloud registration method and system based on deep learning unsupervised |
CN116340540A (en) * | 2023-03-24 | 2023-06-27 | 南京理工大学 | Method for generating network security emergency response knowledge graph based on text |
CN116312484A (en) * | 2023-05-18 | 2023-06-23 | 南京邮电大学 | Cross-language domain invariant acoustic feature extraction method and system |
CN117668256A (en) * | 2023-12-12 | 2024-03-08 | 大连理工大学 | Knowledge graph completion method based on global and local contrast learning |
Non-Patent Citations (2)
Title |
---|
SHUANG LIANG: "Knowledge Graph Embedding Based on Graph Neural Network", 《2023 IEEE 39TH INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE)》, 26 July 2023 (2023-07-26) * |
黄培馨;赵翔;方阳;朱慧明;肖卫东;: "融合对抗训练的端到端知识三元组联合抽取", 计算机研究与发展, no. 12, 15 December 2019 (2019-12-15) * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111628970B (en) | DGA type botnet detection method, medium and electronic equipment | |
CN115170934B (en) | Image segmentation method, system, equipment and storage medium | |
CN112767106B (en) | Automatic auditing method, system, computer readable storage medium and auditing equipment | |
CN115170449B (en) | Multi-mode fusion scene graph generation method, system, equipment and medium | |
CN113723070B (en) | Text similarity model training method, text similarity detection method and device | |
CN111639230B (en) | Similar video screening method, device, equipment and storage medium | |
CN112818162A (en) | Image retrieval method, image retrieval device, storage medium and electronic equipment | |
CN114548300B (en) | Method and device for explaining service processing result of service processing model | |
CN112420125A (en) | Molecular attribute prediction method and device, intelligent equipment and terminal | |
CN115344805A (en) | Material auditing method, computing equipment and storage medium | |
Wei et al. | Efficient local histogram searching via bitmap indexing | |
CN111079930A (en) | Method and device for determining quality parameters of data set and electronic equipment | |
TW202044110A (en) | Unsupervised model evaluation method and device, server and readable storage medium | |
CN110135428A (en) | Image segmentation processing method and device | |
CN111738290A (en) | Image detection method, model construction and training method, device, equipment and medium | |
CN115049852B (en) | Bearing fault diagnosis method and device, storage medium and electronic equipment | |
Canchila et al. | Hyperparameter optimization and importance ranking in deep learning–based crack segmentation | |
CN118036732A (en) | Social event pattern relation completion method and system based on critical countermeasure learning | |
CN115565115A (en) | Outfitting intelligent identification method and computer equipment | |
CN112906824B (en) | Vehicle clustering method, system, device and storage medium | |
US11676050B2 (en) | Systems and methods for neighbor frequency aggregation of parametric probability distributions with decision trees using leaf nodes | |
US11244015B1 (en) | Projecting queries into a content item embedding space | |
Lijun et al. | An intuitionistic calculus to complex abnormal event recognition on data streams | |
Wei et al. | Salient object detection based on weighted hypergraph and random walk | |
Li et al. | Power Analysis Attack Based on Lightweight Convolutional Neural Network |
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 |