CN109325129A - A kind of knowledge mapping inference method, electronic equipment, storage medium and system - Google Patents
A kind of knowledge mapping inference method, electronic equipment, storage medium and system Download PDFInfo
- Publication number
- CN109325129A CN109325129A CN201810589588.7A CN201810589588A CN109325129A CN 109325129 A CN109325129 A CN 109325129A CN 201810589588 A CN201810589588 A CN 201810589588A CN 109325129 A CN109325129 A CN 109325129A
- Authority
- CN
- China
- Prior art keywords
- matrix
- entity
- knowledge mapping
- relationship
- path
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of knowledge mapping inference method, and including the triplet sets insertion in knowledge mapping is transformed into low-dimensional vector space, triplet sets include entity pair and set of relationship, and vector matrix includes entity matrix and relational matrix；The relation path between each entity pair is being searched in low-dimensional vector space according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH；Tensor resolution is carried out to relation path according to knowledge mapping reasoning algorithm and calculates the loss function value of tensor resolution；Entity matrix is matched with relational matrix according to decomposition loss function value and relation path.A kind of knowledge mapping inference method of the invention, vector matrix is obtained by the way that the triplet sets in knowledge mapping are transformed into low-dimensional vector, the relation path between each entity pair is being searched in low-dimensional vector space according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH, and tensor resolution is carried out to relation path using knowledge mapping reasoning algorithm, the pairing of final implementation relation, completes entire reasoning process.
Description
Technical field
The present invention relates to knowledge mapping field more particularly to a kind of knowledge mapping inference methods, electronic equipment, storage medium
And system.
Background technique
Knowledge mapping (Knowledge Graph) is used as a kind of knowledge representation method and Db Management Model, in natural language
The fields such as speech processing, question answering, information retrieval have significant application value.Knowledge mapping is the semantic knowledge-base of structuring,
For describing concept and its correlation in physical world with sign format.
Currently, being directed to knowledge mapping reasoning algorithm, proposes three rank tensor resolution methods, be efficiently applied to knowledge mapping and push away
Reason, but current inference method does not consider path relation between entity in reasoning, ignores the transitivity in path between entity, pushes away
Rationality can be restricted centainly.
Summary of the invention
For overcome the deficiencies in the prior art, one of the objects of the present invention is to provide a kind of knowledge mapping inference method,
It can solve current inference method and does not consider path relation between entity in reasoning, ignore the transitivity in path between entity,
Reasoning performance is by certain the problem of restricting.
The second object of the present invention is to provide a kind of electronic equipment, can solve current inference method in reasoning simultaneously
Path relation between entity is not considered, ignores the transitivity in path between entity, reasoning performance is by certain the problem of restricting.
The third object of the present invention is to provide a kind of storage medium, can solve current inference method in reasoning simultaneously
Path relation between entity is not considered, ignores the transitivity in path between entity, reasoning performance is by certain the problem of restricting.
The fourth object of the present invention is to provide a kind of knowledge mapping inference system, can solve current inference method and exist
Path relation between entity is not considered in reasoning, ignores the transitivity in path between entity, what reasoning performance was centainly restricted asks
Topic.
An object of the present invention is implemented with the following technical solutions:
A kind of knowledge mapping inference method, characterized by comprising:
Triplet sets insertion in knowledge mapping is transformed into low-dimensional vector space, the triple collection by space conversion
Conjunction is converted to vector matrix, and the triplet sets include entity pair and set of relationship, and the vector matrix includes entity square
Battle array and relational matrix；
Each institute is searched according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH in search relationship path in the low-dimensional vector space
State the relation path between entity pair；
Tensor resolution carries out tensor resolution to the relation path according to knowledge mapping reasoning algorithm and calculates the tensor
The loss function value of decomposition；
Matrix pairing, according to the decomposition loss function value and the relation path to the entity matrix and the pass
It is that matrix is matched.
It further, further include matrix update, it is corresponding to update the entity in the vector matrix using least square method
Relationship in the set of relationship.
Further, the entity to include head entity to and tail entity pair, the set of relationship be the head entity pair
In the tail entity to set of relationship that may be present, the entity matrix include head entity to matrix and tail entity to matrix.
Further, the search relationship path specifically: real in the head according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH
An initial solid is selected in volume matrix, reaches the initial solid in the tail entity matrix simultaneously by random walk mode
Generate corresponding relation path.
Further, the matrix pairing specifically: existed according to the decomposition loss function value and the relation path
Corresponding head entity is found out in the head entity matrix, corresponding tail entity and the relational matrix in the tail entity matrix
In corresponding relationship and establish mapping relations.
The second object of the present invention is implemented with the following technical solutions:
A kind of electronic equipment, characterized by comprising: processor；
Memory；And program, wherein described program is stored in the memory, and is configured to by processor
It executes, described program includes for executing a kind of knowledge mapping inference method of the invention.
The third object of the present invention is implemented with the following technical solutions:
A kind of computer readable storage medium, is stored thereon with computer program, it is characterised in that: the computer program
It is executed by processor a kind of knowledge mapping inference method of the invention.
The fourth object of the present invention is implemented with the following technical solutions:
A kind of knowledge mapping inference system, characterized by comprising:
Space conversion module, the space conversion module are used to for the triplet sets insertion in knowledge mapping being transformed into low
Dimensional vector space, the triplet sets are converted to vector matrix, and the triplet sets include entity pair and set of relationship,
The vector matrix includes entity matrix and relational matrix；
Search relationship path module, the search relationship path module according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH for existing
The relation path between each entity pair is searched in the low-dimensional vector space；
Tensor resolution module, the tensor resolution module be used for according to knowledge mapping reasoning algorithm to the relation path into
Row tensor resolution and the loss function value for calculating the tensor resolution.
It further, further include matrix matching module, the matrix matching module is used for according to the decomposition loss function
Value and the relation path match the entity matrix with the relational matrix.
Further, the tensor resolution module includes decomposition unit and computing unit, and the decomposition unit is used for basis
Knowledge mapping reasoning algorithm carries out tensor resolution to the relation path, and the computing unit is for calculating the tensor resolution
Loss function value.
Compared with prior art, the beneficial effects of the present invention are a kind of knowledge mapping inference methods of the invention, including sky
Between convert, the triplet sets insertion in knowledge mapping is transformed into low-dimensional vector space, the triplet sets be converted to
Moment matrix, the triplet sets include entity pair and set of relationship, and the vector matrix includes entity matrix and relationship
Matrix；Search relationship path is searched in the low-dimensional vector space each described according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH
Relation path between entity pair；Tensor resolution carries out tensor resolution to the relation path according to knowledge mapping reasoning algorithm
And calculate the loss function value of the tensor resolution；Matrix pairing, according to the decomposition loss function value and the relationship road
Diameter matches the entity matrix with the relational matrix；By the way that the triplet sets in knowledge mapping are transformed into low-dimensional
Vector matrix is obtained in vector, and each institute is searched in the low-dimensional vector space according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH
The relation path between entity pair is stated, and tensor resolution is carried out to relation path using knowledge mapping reasoning algorithm, it is final to realize
Entire reasoning process is completed in the pairing of relationship.
The above description is only an overview of the technical scheme of the present invention, in order to better understand the technical means of the present invention,
And can be implemented in accordance with the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention and the accompanying drawings.
A specific embodiment of the invention is shown in detail by following embodiment and its attached drawing.
Detailed description of the invention
The drawings described herein are used to provide a further understanding of the present invention, constitutes part of this application, this hair
Bright illustrative embodiments and their description are used to explain the present invention, and are not constituted improper limitations of the present invention.In the accompanying drawings:
Fig. 1 is a kind of flow chart of knowledge mapping inference method of the invention；
Fig. 2 is a kind of module frame chart of knowledge mapping inference system of the invention.
Specific embodiment
In the following, being described further in conjunction with attached drawing and specific embodiment to the present invention, it should be noted that not
Under the premise of conflicting, new implementation can be formed between various embodiments described below or between each technical characteristic in any combination
Example.In the present embodiment, knowledge mapping (Knowledge Graph) is used as a kind of knowledge representation method and Db Management Model,
There is significant application value in fields such as natural language processing, question answering, information retrievals.Knowledge mapping is the semanteme of structuring
Knowledge base, for describing concept and its correlation in physical world with sign format.It mainly uses triple form
(head, relation, tail) carries out the representation of knowledge, and head is head entity, and tail is tail entity, and relation is relationship, real
It is interconnected between body by relationship, the webbed structure of knowledge of shape.
As shown in Figure 1, a kind of knowledge mapping inference method of the invention, specifically includes the following steps:
Triplet sets insertion in knowledge mapping is transformed into low-dimensional vector space by space conversion, and triplet sets turn
It is changed to vector matrix, triplet sets include entity pair and set of relationship, and vector matrix includes entity matrix and relationship square
Battle array.Entity to include head entity to and tail entity pair, set of relationship be head entity for tail entity to set of relations that may be present
Close, entity matrix include head entity to matrix and tail entity to matrix.
Each entity pair is searched according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH in search relationship path in low-dimensional vector space
Between relation path.Search relationship path specifically: selected in head entity matrix according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH
An initial solid is selected, reach initial solid in tail entity matrix by random walk mode and generates corresponding relationship road
Diameter.
Tensor resolution carries out tensor resolution to relation path according to knowledge mapping reasoning algorithm and calculates the damage of tensor resolution
Lose functional value.
Matrix pairing, matches entity matrix with relational matrix according to decomposition loss function value and relation path.
It further include matrix update, using the pass in the corresponding set of relationship of entity in least square method renewal vector matrix
System.Corresponding head entity is found out in head entity matrix, in tail entity matrix according to loss function value and relation path is decomposed
Corresponding relationship and mapping relations are established in corresponding tail entity and relational matrix.
The present invention provides a kind of electronic equipment, comprising: processor；
Memory；And program, wherein program is stored in memory, and is configured to be executed by processor, journey
Sequence includes for executing a kind of knowledge mapping inference method of the invention.
The present invention provides a kind of computer readable storage medium, is stored thereon with computer program, and computer program is located
Reason device executes a kind of knowledge mapping inference method of the invention.
As shown in Fig. 2, the present invention also provides a kind of knowledge mapping inference systems, comprising:
Space conversion module, space conversion module be used for by knowledge mapping triplet sets insertion be transformed into low-dimensional to
Quantity space, triplet sets are converted to vector matrix, and triplet sets include entity pair and set of relationship, and vector matrix includes
Entity matrix and relational matrix；
Search relationship path module, search relationship path module are used for according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH in low-dimensional
The relation path between each entity pair is searched in vector space；
Tensor resolution module, tensor resolution module are used to carry out tensor point to relation path according to knowledge mapping reasoning algorithm
Solve and calculate the loss function value of tensor resolution.Including matrix matching module, matrix matching module is used to lose letter according to decomposition
Numerical value and relation path match entity matrix with relational matrix.Tensor resolution module includes decomposition unit and calculates single
Member, decomposition unit are used to carry out tensor resolution to relation path according to knowledge mapping reasoning algorithm, and computing unit is opened for calculating
Measure the loss function value decomposed.
A kind of knowledge mapping inference method of the invention, including space conversion, the triplet sets in knowledge mapping are embedding
Enter to be transformed into low-dimensional vector space, triplet sets are converted to vector matrix, and triplet sets include entity pair and set of relations
It closes, vector matrix includes entity matrix and relational matrix；Search relationship path exists according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH
The relation path between each entity pair is searched in low-dimensional vector space；Tensor resolution, according to knowledge mapping reasoning algorithm to pass
It is that path carries out tensor resolution and calculates the loss function value of tensor resolution；Matrix pairing, according to decompose loss function value and
Relation path matches entity matrix with relational matrix；By by the triplet sets in knowledge mapping be transformed into low-dimensional to
Vector matrix is obtained in amount, and each entity is searched to it in low-dimensional vector space according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH
Between relation path, and using knowledge mapping reasoning algorithm to relation path carry out tensor resolution, the pairing of final implementation relation,
Complete entire reasoning process.
More than, only presently preferred embodiments of the present invention is not intended to limit the present invention in any form；All current rows
The those of ordinary skill of industry can be shown in by specification attached drawing and above and swimmingly implement the present invention；But all to be familiar with sheet special
The technical staff of industry without departing from the scope of the present invention, is made a little using disclosed above technology contents
The equivalent variations of variation, modification and evolution is equivalent embodiment of the invention；Meanwhile all substantial technologicals according to the present invention
The variation, modification and evolution etc. of any equivalent variations to the above embodiments, still fall within technical solution of the present invention
Within protection scope.
Claims (10)
1. a kind of knowledge mapping inference method, characterized by comprising:
Triplet sets insertion in knowledge mapping is transformed into low-dimensional vector space by space conversion, and the triplet sets turn
Be changed to vector matrix, the triplet sets include entity pair and set of relationship, the vector matrix include entity matrix with
And relational matrix；
Each reality is searched according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH in search relationship path in the low-dimensional vector space
Relation path between body pair；
Tensor resolution carries out tensor resolution to the relation path according to knowledge mapping reasoning algorithm and calculates the tensor resolution
Loss function value；
Matrix pairing, according to the decomposition loss function value and the relation path to the entity matrix and the relationship square
Battle array is matched.
2. a kind of knowledge mapping inference method as described in claim 1, it is characterised in that: further include matrix update, using most
Small square law updates the relationship in the corresponding set of relationship of the entity in the vector matrix.
3. a kind of knowledge mapping inference method as described in claim 1, it is characterised in that: the entity is to including head entity pair
With tail entity pair, the set of relationship is for the head entity for the tail entity to set of relationship that may be present, the reality
Volume matrix include head entity to matrix and tail entity to matrix.
4. a kind of knowledge mapping inference method as claimed in claim 3, it is characterised in that: the search relationship path is specific
Are as follows: an initial solid is selected in the head entity matrix according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH, passes through random walk
Mode reaches the initial solid in the tail entity matrix and generates corresponding relation path.
5. a kind of knowledge mapping inference method as claimed in claim 3, it is characterised in that: the matrix pairing specifically: root
Corresponding head entity, the tail are found out in the head entity matrix according to the decomposition loss function value and the relation path
Corresponding relationship and mapping relations are established in corresponding tail entity and the relational matrix in entity matrix.
6. a kind of electronic equipment, characterized by comprising: processor；
Memory；And program, wherein described program is stored in the memory, and is configured to be held by processor
Row, described program include requiring method described in 1-5 any one for perform claim.
7. a kind of computer readable storage medium, is stored thereon with computer program, it is characterised in that: the computer program quilt
Processor executes the method as described in claim 1-5 any one.
8. a kind of knowledge mapping inference system, characterized by comprising:
Space conversion module, the space conversion module be used for by knowledge mapping triplet sets insertion be transformed into low-dimensional to
Quantity space, the triplet sets are converted to vector matrix, and the triplet sets include entity pair and set of relationship, described
Vector matrix includes entity matrix and relational matrix；
Search relationship path module, the search relationship path module are used for according to propagation type HEURISTIC ALGORITHM FOR GRAPH SEARCH described
The relation path between each entity pair is searched in low-dimensional vector space；
Tensor resolution module, the tensor resolution module is for opening the relation path according to knowledge mapping reasoning algorithm
Amount decomposes and calculates the loss function value of the tensor resolution.
9. a kind of knowledge mapping inference system as claimed in claim 8, it is characterised in that: further include matrix matching module, institute
State matrix matching module for according to the decomposition loss function value and the relation path to the entity matrix with it is described
Relational matrix is matched.
10. a kind of knowledge mapping inference system as claimed in claim 8, it is characterised in that: the tensor resolution module includes
Decomposition unit and computing unit, the decomposition unit are used to carry out tensor to the relation path according to knowledge mapping reasoning algorithm
It decomposes, the computing unit is used to calculate the loss function value of the tensor resolution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810589588.7A CN109325129A (en) | 2018-06-08 | 2018-06-08 | A kind of knowledge mapping inference method, electronic equipment, storage medium and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810589588.7A CN109325129A (en) | 2018-06-08 | 2018-06-08 | A kind of knowledge mapping inference method, electronic equipment, storage medium and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109325129A true CN109325129A (en) | 2019-02-12 |
Family
ID=65263447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810589588.7A Pending CN109325129A (en) | 2018-06-08 | 2018-06-08 | A kind of knowledge mapping inference method, electronic equipment, storage medium and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109325129A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109885698A (en) * | 2019-02-13 | 2019-06-14 | 北京航空航天大学 | A kind of knowledge mapping construction method and device, electronic equipment |
| CN110309154A (en) * | 2019-06-28 | 2019-10-08 | 京东数字科技控股有限公司 | Substance feature selection method, device, equipment and storage medium based on map |
| CN110491106A (en) * | 2019-07-22 | 2019-11-22 | 深圳壹账通智能科技有限公司 | Data early warning method, device and the computer equipment of knowledge based map |
| CN110796254A (en) * | 2019-10-30 | 2020-02-14 | 南京工业大学 | Knowledge graph reasoning method and device, computer equipment and storage medium |
| CN111739657A (en) * | 2020-07-20 | 2020-10-02 | 北京梦天门科技股份有限公司 | Epidemic infected person prediction method and system based on knowledge graph |
| CN111897972A (en) * | 2020-08-06 | 2020-11-06 | 南方电网科学研究院有限责任公司 | Data track visualization method and device |
| CN112579795A (en) * | 2020-12-28 | 2021-03-30 | 重庆邮电大学 | Intelligent question-answering method based on knowledge graph embedded representation |
| CN118133963A (en) * | 2024-05-07 | 2024-06-04 | 广东南方电信规划咨询设计院有限公司 | Knowledge subgraph construction method and device |
-
2018
- 2018-06-08 CN CN201810589588.7A patent/CN109325129A/en active Pending
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109885698A (en) * | 2019-02-13 | 2019-06-14 | 北京航空航天大学 | A kind of knowledge mapping construction method and device, electronic equipment |
| CN110309154A (en) * | 2019-06-28 | 2019-10-08 | 京东数字科技控股有限公司 | Substance feature selection method, device, equipment and storage medium based on map |
| CN110309154B (en) * | 2019-06-28 | 2021-06-29 | 京东数字科技控股有限公司 | Entity feature selection method, device and equipment based on map and storage medium |
| CN110491106B (en) * | 2019-07-22 | 2022-03-18 | 深圳壹账通智能科技有限公司 | Data early warning method and device based on knowledge graph and computer equipment |
| CN110491106A (en) * | 2019-07-22 | 2019-11-22 | 深圳壹账通智能科技有限公司 | Data early warning method, device and the computer equipment of knowledge based map |
| CN110796254A (en) * | 2019-10-30 | 2020-02-14 | 南京工业大学 | Knowledge graph reasoning method and device, computer equipment and storage medium |
| CN110796254B (en) * | 2019-10-30 | 2024-02-27 | 南京工业大学 | Knowledge graph reasoning method and device, computer equipment and storage medium |
| CN111739657A (en) * | 2020-07-20 | 2020-10-02 | 北京梦天门科技股份有限公司 | Epidemic infected person prediction method and system based on knowledge graph |
| CN111897972B (en) * | 2020-08-06 | 2023-10-17 | 南方电网科学研究院有限责任公司 | Data track visualization method and device |
| CN111897972A (en) * | 2020-08-06 | 2020-11-06 | 南方电网科学研究院有限责任公司 | Data track visualization method and device |
| CN112579795A (en) * | 2020-12-28 | 2021-03-30 | 重庆邮电大学 | Intelligent question-answering method based on knowledge graph embedded representation |
| CN118133963A (en) * | 2024-05-07 | 2024-06-04 | 广东南方电信规划咨询设计院有限公司 | Knowledge subgraph construction method and device |
| CN118133963B (en) * | 2024-05-07 | 2024-07-05 | 广东南方电信规划咨询设计院有限公司 | Knowledge subgraph construction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109325129A (en) | A kind of knowledge mapping inference method, electronic equipment, storage medium and system | |
| CN108009285B (en) | Forest Ecology man-machine interaction method based on natural language processing | |
| CN111143536B (en) | Information extraction method based on artificial intelligence, storage medium and related device | |
| CN111274790B (en) | Chapter-level event embedding method and device based on syntactic dependency graph | |
| CN110580288A (en) | text classification method and device based on artificial intelligence | |
| CN112765312A (en) | Knowledge graph question-answering method and system based on graph neural network embedding matching | |
| CN111966787B (en) | Knowledge graph-based construction method of intelligent fishery question-answering robot | |
| CN116932776A (en) | Knowledge graph-based large model knowledge updating method and device | |
| CN117473053A (en) | Natural language question-answering method, device, medium and equipment based on large language model | |
| CN103679715B (en) | A kind of handset image feature extracting method based on Non-negative Matrix Factorization | |
| CN115222048A (en) | Training method, device, equipment and medium for document abstract generation model | |
| CN113761220A (en) | Information acquisition method, device, equipment and storage medium | |
| CN111598252A (en) | University computer basic knowledge problem solving method based on deep learning | |
| CN116861001A (en) | Medical common sense knowledge graph automatic construction method based on meta learning | |
| CN109685208A (en) | A kind of method and device accelerated for the dilute combization of neural network processor data | |
| CN106779090B (en) | A kind of feature learning model based on adaptive Dropout Non-negative Matrix Factorization | |
| CN112579795A (en) | Intelligent question-answering method based on knowledge graph embedded representation | |
| CN108256030A (en) | A kind of degree adaptive Concept Semantic Similarity computational methods based on ontology | |
| CN116051961A (en) | Target detection model training method, target detection method, device and medium | |
| CN114611990A (en) | Method and device for evaluating contribution rate of element system of network information system | |
| CN103279578A (en) | Video retrieving method based on context space | |
| Sun et al. | Uniform persistence and periodic solutions of generalized predator–prey type eco-epidemiological systems | |
| CN106650777B (en) | In conjunction with the Target Segmentation method of dictionary group | |
| CN111931478B (en) | Training method of address interest surface model, and prediction method and device of address | |
| Chen et al. | Analysis of a stochastic nonautonomous hybrid population model with impulsive perturbations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190212 |
|
| WD01 | Invention patent application deemed withdrawn after publication |