CN106934042A - A kind of knowledge mapping represents model and its method - Google Patents
A kind of knowledge mapping represents model and its method Download PDFInfo
- Publication number
- CN106934042A CN106934042A CN201710155940.1A CN201710155940A CN106934042A CN 106934042 A CN106934042 A CN 106934042A CN 201710155940 A CN201710155940 A CN 201710155940A CN 106934042 A CN106934042 A CN 106934042A
- Authority
- CN
- China
- Prior art keywords
- entity
- module
- model
- space
- knowledge
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000013507 mapping Methods 0.000 title abstract description 14
- 239000013598 vector Substances 0.000 claims abstract description 69
- 238000012549 training Methods 0.000 claims abstract description 33
- 238000013519 translation Methods 0.000 claims abstract description 15
- 239000011159 matrix material Substances 0.000 claims description 15
- 238000001914 filtration Methods 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 10
- 238000010276 construction Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 abstract description 3
- 230000006870 function Effects 0.000 description 17
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 4
- 239000010931 gold Substances 0.000 description 4
- 229910052737 gold Inorganic materials 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 240000003421 Dianthus chinensis Species 0.000 description 1
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000009828 non-uniform distribution Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Model and its method are represented the invention discloses a kind of knowledge mapping, it is related to knowledge mapping presentation technology field, the expression model includes entity space module, majorized function module and model training module；The entity space is used for the representation space of presentation-entity feature, and it includes intrinsic state space and mimicry space；The majorized function is used to represent different entities distance after translation that it to include that distance is calculated and weight vectors；The model training module is used for features training and exports training result, and the training result is used to carry out knowledge mapping prediction and classification.Translation model of the present invention based on dynamic representation space, one dynamic representation space is set to each relation, solve the problems, such as to represent that model cannot be distinguished by different relation spaces in the prior art, there is provided more fully reliable method for expressing, and the efficiency that the complexity of model improves algorithm is reduced, preferable effect is achieved in actual applications.
Description
Technical Field
The invention relates to the technical field of knowledge graph representation, in particular to a knowledge graph representation model and a method thereof.
Background
The existing knowledge graph representation methods around the world mainly focus on utilizing artificially constructed features and features represented based on an RDF framework. These feature representation methods have problems of low efficiency, complex algorithm, and the like in knowledge representation. In recent years, a series of knowledge representation methods are proposed by using a deep learning method, but the current training knowledge representation method has the problems of higher model complexity or lower training efficiency.
In addition, representative work of a knowledge graph representation method at home and abroad mainly comprises TransE (embedded model based on translation) [1]]TransH (Hyperplanar-based Embedded model) [2]]TransR (Embedded model based on entity relationship space) [3]CTRANsR (Embedded model based on clustering and entity relationship space) [3]And TransD (Embedded model based on dynamic mapping matrix) [4]The methods are collectively referred to as translation-based knowledge representation models. Based on the translation model, for each triplet (h, r, t), the relation r is a translation operation from the head entity vector h to the tail entity vector t, and accordingly, Bordes et al first proposed a TransE (embedded model based on translation) knowledge representation method, which measures semantic similarity between calculated entities by the offset in Euclidean distance, and is a simple and basic knowledge representation methodr(h,t)＝||h+r-t||2Wherein | | h + r-t | | non-woven phosphor2Is the norm of order 2 of h + r-t, the euclidean distance. The TransH (embedded model based on hyperplane) method establishes a relation-oriented hyperplane which is composed of a normal vector nrAnd a translation vector r, the head entity vector h and the tail entity vector t are firstly projected to a hyperplane of the relation to obtain a vector h⊥＝h-nr ThnrAnd t⊥＝t-nr Thnr. thus, the optimization goal of TransH (Hyperplanar-based Embedded model) becomes h⊥+r＝t⊥With the corresponding score function modified to fr(h,t)＝||h⊥+r-t⊥||2. TransR (Embedded model based on entity relationship space) and CTRANsR (Embedded model based on clustering and entity relationship space) are expected to build a mapping matrix MrAnd a vector r for representing each relation r, specifically, a head entity vector h and a tail entity vector t are mapped to the hierarchy of the relation vector r through a matrix by TransR (embedded model based on entity relation space), and M is obtainedrh+r＝Mrt, namely the optimization target of the transR (embedded model based on the entity relationship space), the TransD (embedded model based on the dynamic mapping matrix) replaces the multiplication operation of the matrix and the vector in the transR (embedded model based on the entity relationship space) by the vector operation, and the algorithm efficiency is improved.
In practical application, TransE (embedded model based on translation) [1] achieves better prediction effect. In TransE (a translation-based embedded model), for each triplet (h, r, t), the head entity vector h, the tail entity vector t, and the relationship r are represented as n-dimensional vectors h (t) and r. The embedded vector t is approximately equal to the embedded h plus the embedded r, i.e. h + r ≈ t, and the TransE (translation-based embedded model) can well handle one-to-one relationships, but has an obvious disadvantage in handling complex relationships such as one-to-N, N-to-one, and N-to-N, and particularly, in the process of handling complex relationships, different entities use the same vector, which is not practical. TransH (hyperplane-based embedded model) [2] solves the problem of complex relationships by a hyperplane mapping rule that maps head entity vectors h and tail entity vectors t to relationship-specific hyperplanes. But entities and relationships are two completely different concepts and therefore it is not correct to place them in the same vector space. TransR (Embedded model based on entity relationship space)/CTRANsR (Embedded model based on clustering and entity relationship space) [3] and TransD (Embedded model based on dynamic mapping matrix) [4] propose two novel models that put entities and relationships in different vector spaces, for example: entity space and multiple relation space, TransR (embedded model based on entity relation space) sets a mapping matrix Mr for each relation r, and then the entity is mapped into the relation space by using Mr. In the relationship space, the entity vector mapped by Mr and the relationship vector r can construct a gold triple, which is described as Mrh + r ≈ Mrt. As an extension to TransR (Embedded model based on entity relationship space), CTRANsR (Embedded model based on clustering and entity relationship space) uses a clustering algorithm to segment the initial results of TransE (Embedded model based on translation), dividing each relationship r into several sub-relationships rs. To some extent, replacing r with rs solves the ambiguity problem for each relationship. TransD (embedded model based on dynamic mapping matrix) constructs a dynamic mapping matrix for each entity-relationship pair using two vectors ep and hp. However, the algorithm complexity of transR (embedded model based on entity relationship space)/CTRANsR (embedded model based on clustering and entity relationship space) is high, and the method cannot be applied in practice.
【1】Bordes A,Usunier N,Garcia-Duran A,et al.Translating embeddings formodeling multi-relational data[C]//Proc of NIPS.Cambridge,MA:MIT Press,2013:2787–2795
【2】Wang Zhen,Zhang Jianwen,Feng Jianlin,et al.Knowledge graphembedding by translating on hyperplanes[C]//Proc of AAAI.Menlo Park,CA:AAAI,2014:1112–1119
【3】Lin Yankai,Liu Zhiyuan,Sun Maosong,et al.Learning entity andrelation embeddings for knowledge graph completion[C]//Proc of AAAI.MenloPark,CA:AAAI,2015
【4】Ji Guoliang,He Shizhu,Xu Liheng,et al.Knowledge graph embeddingvia dynamic mapping matrix[C]//Proc of ACL.Stroudsburg PA:ACL,2045:687–696
Disclosure of Invention
The invention aims to provide a knowledge graph representation model and a method thereof, which can improve the representation capability of a knowledge graph and have important functions on completion and verification of the knowledge graph.
In order to achieve the purpose of the present invention, a first technical solution of the present invention is specifically as follows:
a knowledge spectrogram representation model comprises an entity space module, an optimization function module and a model training module;
the entity space module is used for representing a representation space of entity features, and comprises an eigen state space and a mimicry state space;
the optimization function module is used for representing the distance of different entities after translation, and comprises distance calculation and weight vectors;
the model training module is used for feature training and outputting a training result, and the training result is used for predicting and classifying the knowledge graph.
The second technical scheme of the invention is as follows:
a method of constructing the knowledge spectrogram representation model, the method comprising the steps of:
1) an entity space construction process, which adopts two vectors to represent an entity and a relation, wherein the two vectors comprise an eigen state vector and a mimicry vector, the eigen state vector is used for describing an entity relation eigen state, the mimicry vector is used for describing an entity relation mimicry, the mimicry vector forms a mimicry matrix, and the mimicry vector and the eigen state vector jointly form a feature vector of an entity space;
2) the optimization function process comprises the steps of calculating a distance formula between the translated head entity and the translated tail entity, and expressing weights of different dimensions by adopting weight vectors so as to achieve the purpose of optimizing the distance calculation formula;
3) a model training process that includes dynamic training of weight vectors and setting of parameters that prevent overfitting.
The third technical scheme of the invention is as follows:
the implementation method of the knowledge spectrogram representation model further comprises a data acquisition module, a preprocessing module, a feature extraction module, a training module, a knowledge spectrogram completion module and a classification module, and specifically comprises the following steps:
1) extracting data in the existing knowledge map by using a data acquisition module, performing distributed acquisition on knowledge existing in the Internet by using a distributed crawler system, and storing the knowledge in a distributed map database;
2) the method comprises the steps that a preprocessing module is used for carrying out structuralized processing on extracted data, and the preprocessing module is used for filtering the collected data and mainly comprises three parts, namely, entity relation duplicate removal, filtering entity relations which do not accord with description specifications and filtering entity relations in which illegal characters exist;
3) performing feature extraction on the data after the structured processing by using a feature extraction module, extracting entities, relations and attributes contained in the knowledge graph, describing the entities, relations and attributes in a triple form, and training the extracted features by using a training module;
4) and predicting and classifying the knowledge graph through the knowledge graph complementing module and the classifying module by using the trained result, wherein the knowledge graph complementing module and the classifying module test the trained representation model to verify the effectiveness of the model, and the recommendation of missing entities or relations in the knowledge graph and the judgment of the correctness of the existing triples are realized.
Compared with the prior art, the invention has the following beneficial effects:
1. the invention sets a dynamic representation space for each relation based on the translation model of the dynamic representation space, solves the problem that the representation model can not distinguish different relation spaces in the prior art, provides a more comprehensive and reliable representation method, reduces the complexity of the model, improves the efficiency of the algorithm, and achieves better effect in practical application.
2. The invention provides a novel TransDR model (an embedded model based on a dynamic relation space), which constructs a dynamic relation space for each relation and provides a self-adaptive relation weight for each relation vector space; TransDR (an embedded model based on dynamic relationship space) can reduce noise from other relationships and improve the discrimination between different relationships.
Drawings
FIG. 1 is a schematic diagram of the composition and structure of a knowledge spectrogram representation model in the present invention.
Fig. 2 is an explanatory diagram of the basic idea of the TransDR model in the present invention.
FIG. 3 is a flow chart of entity space construction of a knowledge graph representation model in the present invention.
FIG. 4 is a flow chart of the optimization function construction of the knowledge graph representation model in the present invention.
FIG. 5 is a flow chart of the knowledge graph representation model training of the present invention.
FIG. 6 is a flow chart of the knowledge graph representation model operation of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
The embodiment of the invention relates to a knowledge graph representation model and a method thereof, and the knowledge graph representation model comprises a data acquisition module, a preprocessing module, a feature extraction module, a knowledge graph completion module and a classification module, as shown in fig. 1.
The data acquisition module acquires knowledge existing in the Internet in a distributed manner by using a distributed crawler system and stores the knowledge in a distributed graph database; the preprocessing module filters the acquired data and mainly comprises three parts, namely, entity relationship duplication removal, entity relationship filtering, and entity relationship filtering, wherein the entity relationship filtering is not in accordance with the description specification, and illegal characters exist; the characteristic extraction module is used for training the extracted entity through a knowledge graph representation model; the knowledge graph complementing module and the classification module test the trained representation model, and the effectiveness of the model is verified.
The training module is a knowledge graph representation model, and the establishment process is as follows:
definition G is a triplet (h, r, t), where h represents the head entity, r represents the relationship, t represents the tail entity, and the bold font word h, r, t is the embedded representation of (h, r, t).
h＝(h1，h2，...，hi，...，hn)；
r＝(r1，r2，...，ri，...，rn)；
t＝(t1，t2，...，ti，...，tn)；
hi，ri，tiThe ith feature is h, r, t, respectively, n is the length of the feature, Δ represents the correct triplet set, Δ ' represents the incorrect triplet set, so (h, r, t) ∈ Δ represents that the triplet is correct, (h ', r ', t ') ∈ Δ ' represents that the triplet is incorrect, WrRepresenting an embedded representation of the relation weights, Wr＝(wr1，wr2，...，wri，...，wrn)；
The construction of the knowledge graph representation model specifically comprises entity space construction, optimization function and model training.
Solid space construction
Since the distribution of potential features of the knowledge-graph in different dimensions is not uniform, equal weights cannot be set for each dimension. The invention thus proposes a more flexible model, namely the TransDR (Embedded model based on dynamic relational space), which takes into account different types of relational spaces to avoid uneven distribution of different dimensions. As shown in fig. 2, the nsdr (embedded model based on dynamic relationship space) defines two vectors for each entity h (t), three vectors for each relationship r, a first vector of h (t) and a first vector of r represent natural features, a second vector of h (t) and a second vector of r represent extrinsic features used to construct the mapping matrix, and a third vector of r represents the relationship space weight in each dimension. Entities are randomly unordered in entity space, but after being translated into relationship space, the entities become ordered. Although it is used for
i＝hi⊥+r-ti⊥，WhereiniThe distance between the correct triples is represented,indicates the distance between the correct triplets, but r1(r2) The space can effectively identify the wrong tail entity and the correct tail entity, as shown in fig. 2 in particular.
The invention uses two vectors to represent entities and relationships, the first vector describing the eigenstates of the entity relationships, the other vector describing the mimicry of the entity relationships, six vectors he，hm，re，rm，te,tmA triplet (h, r, t) is represented, where the subscripts e and m denote the eigenstates and mimicry, the true state of an entity being composed of both mimicry and eigenstates.
Wherein h ise,hm,re,rm,te,tm∈Gm。
RmRepresenting an M-dimensional entity relationship space, for each triplet (h, r, t), the invention defines two mimicry matrices Mh、MtRepresenting head and tail entity mimicry spaces, respectively, for Mh、MtIs defined as follows:
Mh＝rmhm T(1)
Mt＝rmtm T(2)
the symmetric mimicry matrix is constructed from a mimicry vector of entities and relationships, so for each triplet (h, r, t), h and t have a unique mimicry matrix.
In fact, the true state of an entity is composed of both mimicry and eigenstates, so the present invention defines the true vector as follows:
h＝Mhhe+he(3)
t＝Mtte+te(4)
the process of constructing the solid-space model is specifically shown in fig. 3.
Optimization function
Scoring function f for a gold tripler(h, t) should be lower and higher for an incorrect gold triplet, denoted L2For norm as an example, the conventional scoring function is:
to solve the problem that the traditional scoring function calculates the Euclidean Distance (ED) between h + r and t, which considers the uniform distribution of each feature dimension, but which cannot distinguish the trends of different relationship spaces, the invention uses a normalized euclidean distance (SED) instead of ED.
Wherein the function of SED is:
where X and X denote the feature vectors before and after normalization, respectively, and μ and σ are the expectation and standard deviation vectors, respectively. SED can eliminate the non-uniform distribution of features by normalization, so the distance between h + r and t is:
wherein σiIs the ith dimension characteristic of sigma, take Wr＝1/σ2Book, bookThe invention proposes a new scoring function:
the optimization function construction flow is specifically shown in fig. 4.
Model training
To obtain a distinction between gold triples and incorrect triples, the present invention defines the following margin-based ranking loss function:
wherein,Δ and Δ' are the sets of positive and negative triplets, respectively, and γ is the distance between positive and negative samples. Since the original triple set contains only positive triples, negative triples have to be generated manually.
The invention constructs a negative triple set by using two strategies of random sampling and Bernoulli sampling.
The following constraints are considered in minimizing the loss L:
||h||≤1，||t||≤1
||Wr||＝1 (7)
where equation (7) ensures that the relationship weight vector Wr is a constant, it is converted by means of soft constraints into the following unconstrained loss function:
where λ and η are two hyper-parameters that weight the importance of the soft constraint.
Training the penalty function using modified random gradient descent (ADADELTA), with constraints (5) and (6) missing in equation (8), and instead, constraints (5) and (6) are satisfied directly before each mini-lot is visited; to speed up convergence, the result of Trans E is used instead of he，te，reThe flow of the model training is specifically shown in fig. 5.
The working process of the knowledge graph representation model is shown in fig. 6, and the implementation method of the knowledge graph representation model comprises the following steps:
1) extracting data in the existing knowledge map by using a data acquisition module, performing distributed acquisition on knowledge existing in the Internet by using a distributed crawler system, and storing the knowledge in a distributed map database;
2) the method comprises the steps that a preprocessing module is used for carrying out structuralized processing on extracted data, and the preprocessing module is used for filtering the collected data and mainly comprises three parts, namely, entity relation duplicate removal, filtering entity relations which do not accord with description specifications and filtering entity relations in which illegal characters exist;
3) performing feature extraction on the data after the structured processing by using a feature extraction module, extracting entities, relations and attributes contained in the knowledge graph, describing the entities, relations and attributes in a triple form, and training the extracted features by using the knowledge graph representation model;
4) and predicting and classifying the knowledge graph through the knowledge graph complementing module and the classifying module by using the trained result, wherein the knowledge graph complementing module and the classifying module test the trained representation model to verify the effectiveness of the model, and the recommendation of missing entities or relations in the knowledge graph and the judgment of the correctness of the existing triples are realized.
The above embodiments only serve to explain the technical solution of the present invention, and the protection scope of the present invention is not limited to the implementation system and the specific implementation steps described in the above embodiments. Therefore, the technical solutions that the specific formulas and algorithms in the above embodiments are simply replaced, but the substantial contents are still consistent with the method of the present invention, and all the technical solutions are within the protection scope of the present invention.
Claims (3)
1. A knowledge spectrogram representation model is characterized by comprising an entity space module, an optimization function module and a model training module;
the entity space module is used for representing a representation space of entity features, and comprises an eigen state space and a mimicry state space;
the optimization function module is used for representing the distance of different entities after translation, and comprises distance calculation and weight vectors;
the model training module is used for feature training and outputting a training result, and the training result is used for predicting and classifying the knowledge graph.
2. A method of constructing a knowledge spectrogram representation model of claim 1, the method comprising:
1) an entity space construction process, which adopts two vectors to represent an entity and a relation, wherein the two vectors comprise an eigen state vector and a mimicry vector, the eigen state vector is used for describing an entity relation eigen state, the mimicry vector is used for describing an entity relation mimicry, the mimicry vector forms a mimicry matrix, and the mimicry vector and the eigen state vector jointly form a feature vector of an entity space;
2) the optimization function process comprises the steps of calculating a distance formula between the translated head entity and the translated tail entity, and endowing different dimensions of the head entity and the tail entity with different weights by adopting weight vectors so as to achieve the purpose of optimizing the distance calculation formula;
3) a model training process that includes dynamic training of weight vectors and setting of parameters that prevent overfitting.
3. The method for implementing a knowledge spectrogram representation model as defined in claim 1, further comprising a data acquisition module, a preprocessing module, a feature extraction module, a knowledge spectrogram completion module and a classification module, wherein the implementation method specifically comprises the following steps:
1) extracting data in the existing knowledge map by using a data acquisition module, performing distributed acquisition on knowledge existing in the Internet by using a distributed crawler system, and storing the knowledge in a distributed map database;
2) the method comprises the steps that a preprocessing module is used for carrying out structuralized processing on extracted data, and the preprocessing module is used for filtering the collected data and mainly comprises three parts, namely, entity relation duplicate removal, filtering entity relations which do not accord with description specifications and filtering entity relations in which illegal characters exist;
3) performing feature extraction on the data after the structured processing by using a feature extraction module, extracting entities, relations and attributes contained in the knowledge graph, describing the entities, relations and attributes in a triple form, and training the extracted features by using the knowledge graph representation model;
4) and predicting and classifying the knowledge graph through the knowledge graph complementing module and the classifying module by using the trained result, wherein the knowledge graph complementing module and the classifying module test the trained representation model to verify the effectiveness of the model, and the recommendation of missing entities or relations in the knowledge graph and the judgment of the correctness of the existing triples are realized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710155940.1A CN106934042B (en) | 2017-03-16 | 2017-03-16 | Knowledge graph representation system and implementation method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710155940.1A CN106934042B (en) | 2017-03-16 | 2017-03-16 | Knowledge graph representation system and implementation method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106934042A true CN106934042A (en) | 2017-07-07 |
| CN106934042B CN106934042B (en) | 2020-05-29 |
Family
ID=59432514
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710155940.1A Active CN106934042B (en) | 2017-03-16 | 2017-03-16 | Knowledge graph representation system and implementation method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106934042B (en) |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107590237A (en) * | 2017-09-11 | 2018-01-16 | 桂林电子科技大学 | A kind of knowledge mapping based on dynamic translation principle represents learning method |
| CN107943874A (en) * | 2017-11-13 | 2018-04-20 | 平安科技（深圳）有限公司 | Knowledge mapping processing method, device, computer equipment and storage medium |
| CN108052683A (en) * | 2018-01-22 | 2018-05-18 | 桂林电子科技大学 | A kind of knowledge mapping based on cosine measurement rule represents learning method |
| CN108376287A (en) * | 2018-03-02 | 2018-08-07 | 复旦大学 | Multi-valued attribute segmenting device based on CN-DBpedia and method |
| CN109063021A (en) * | 2018-07-12 | 2018-12-21 | 浙江大学 | A kind of knowledge mapping distribution representation method for capableing of encoding relation semanteme Diversity structure |
| CN109165337A (en) * | 2018-10-17 | 2019-01-08 | 珠海市智图数研信息技术有限公司 | A kind of method and system of knowledge based map construction bidding field association analysis |
| CN109189937A (en) * | 2018-08-22 | 2019-01-11 | 阿里巴巴集团控股有限公司 | A kind of characteristic relation recommended method and device, a kind of calculating equipment and storage medium |
| CN109213872A (en) * | 2018-09-11 | 2019-01-15 | 中国电子科技集团公司第二十八研究所 | Knowledge based indicates the entity relationship prediction technique and forecasting system of study |
| CN109241290A (en) * | 2017-07-10 | 2019-01-18 | 华东师范大学 | A kind of knowledge mapping complementing method, device and storage medium |
| CN109614501A (en) * | 2018-12-13 | 2019-04-12 | 浙江工商大学 | A kind of industrial hidden danger standardization report method and system of knowledge based map |
| CN109753571A (en) * | 2019-01-11 | 2019-05-14 | 中山大学 | A kind of scene map lower dimensional space embedding grammar based on secondary theme space projection |
| CN109922075A (en) * | 2019-03-22 | 2019-06-21 | 中国南方电网有限责任公司 | Network security knowledge map construction method and apparatus, computer equipment |
| CN110019766A (en) * | 2017-09-25 | 2019-07-16 | 腾讯科技（深圳）有限公司 | Methods of exhibiting, device, mobile terminal and the readable storage medium storing program for executing of knowledge mapping |
| CN110275966A (en) * | 2019-07-01 | 2019-09-24 | 科大讯飞(苏州)科技有限公司 | A kind of Knowledge Extraction Method and device |
| CN111339321A (en) * | 2020-05-18 | 2020-06-26 | 中国人民解放军国防科技大学 | Triple authenticity detection method and device in knowledge graph |
| CN111392538A (en) * | 2020-03-17 | 2020-07-10 | 浙江新再灵科技股份有限公司 | Elevator comprehensive fault early warning method based on multi-dimensional Internet of things atlas big data |
| CN111680109A (en) * | 2020-04-22 | 2020-09-18 | 北京三快在线科技有限公司 | Knowledge graph representation learning model training method and device and electronic equipment |
| CN112380298A (en) * | 2020-12-03 | 2021-02-19 | 成都航天科工大数据研究院有限公司 | Knowledge graph construction method and system for oil and gas pipeline defect repair |
| CN112784059A (en) * | 2021-01-20 | 2021-05-11 | 和美（深圳）信息技术股份有限公司 | Knowledge graph representation learning method and device, electronic device and storage medium |
| CN113033914A (en) * | 2021-04-16 | 2021-06-25 | 哈尔滨工业大学 | Entity and relation prediction method for machining process knowledge graph |
| CN116705338A (en) * | 2023-08-08 | 2023-09-05 | 中国中医科学院中医药信息研究所 | Traditional Chinese medicine multi-mode knowledge graph reasoning method and device based on rules and paths |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7319998B2 (en) * | 2003-11-14 | 2008-01-15 | Universidade De Coimbra | Method and system for supporting symbolic serendipity |
| CN103207380A (en) * | 2013-03-12 | 2013-07-17 | 西安电子科技大学 | Broadband target direction finding method based on two-dimensional frequency domain sparse constraint |
| CN105630901A (en) * | 2015-12-21 | 2016-06-01 | 清华大学 | Knowledge graph representation learning method |
| CN106156365A (en) * | 2016-08-03 | 2016-11-23 | 北京智能管家科技有限公司 | A kind of generation method and device of knowledge mapping |
-
2017
- 2017-03-16 CN CN201710155940.1A patent/CN106934042B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7319998B2 (en) * | 2003-11-14 | 2008-01-15 | Universidade De Coimbra | Method and system for supporting symbolic serendipity |
| CN103207380A (en) * | 2013-03-12 | 2013-07-17 | 西安电子科技大学 | Broadband target direction finding method based on two-dimensional frequency domain sparse constraint |
| CN105630901A (en) * | 2015-12-21 | 2016-06-01 | 清华大学 | Knowledge graph representation learning method |
| CN106156365A (en) * | 2016-08-03 | 2016-11-23 | 北京智能管家科技有限公司 | A kind of generation method and device of knowledge mapping |
Non-Patent Citations (1)
| Title |
|---|
| 刘峤等: ""知识图谱构建技术综述"", 《计算机研究与发展》 * |
Cited By (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241290A (en) * | 2017-07-10 | 2019-01-18 | 华东师范大学 | A kind of knowledge mapping complementing method, device and storage medium |
| CN107590237A (en) * | 2017-09-11 | 2018-01-16 | 桂林电子科技大学 | A kind of knowledge mapping based on dynamic translation principle represents learning method |
| CN107590237B (en) * | 2017-09-11 | 2021-04-30 | 桂林电子科技大学 | Knowledge graph representation learning method based on dynamic translation principle |
| CN110019766B (en) * | 2017-09-25 | 2023-01-13 | 腾讯科技（深圳）有限公司 | Knowledge graph display method and device, mobile terminal and readable storage medium |
| CN110019766A (en) * | 2017-09-25 | 2019-07-16 | 腾讯科技（深圳）有限公司 | Methods of exhibiting, device, mobile terminal and the readable storage medium storing program for executing of knowledge mapping |
| CN107943874A (en) * | 2017-11-13 | 2018-04-20 | 平安科技（深圳）有限公司 | Knowledge mapping processing method, device, computer equipment and storage medium |
| WO2019091019A1 (en) * | 2017-11-13 | 2019-05-16 | 平安科技（深圳）有限公司 | Knowledge graph processing method and device, computer device and computer storage medium |
| CN108052683A (en) * | 2018-01-22 | 2018-05-18 | 桂林电子科技大学 | A kind of knowledge mapping based on cosine measurement rule represents learning method |
| CN108052683B (en) * | 2018-01-22 | 2021-08-03 | 桂林电子科技大学 | Knowledge graph representation learning method based on cosine measurement rule |
| CN108376287A (en) * | 2018-03-02 | 2018-08-07 | 复旦大学 | Multi-valued attribute segmenting device based on CN-DBpedia and method |
| CN109063021A (en) * | 2018-07-12 | 2018-12-21 | 浙江大学 | A kind of knowledge mapping distribution representation method for capableing of encoding relation semanteme Diversity structure |
| CN109063021B (en) * | 2018-07-12 | 2021-10-15 | 浙江大学 | Knowledge graph distributed expression method capable of coding relation semantic diversity structure |
| CN109189937A (en) * | 2018-08-22 | 2019-01-11 | 阿里巴巴集团控股有限公司 | A kind of characteristic relation recommended method and device, a kind of calculating equipment and storage medium |
| US11244232B2 (en) | 2018-08-22 | 2022-02-08 | Advanced New Technologies Co., Ltd. | Feature relationship recommendation method, apparatus, computing device, and storage medium |
| CN109213872A (en) * | 2018-09-11 | 2019-01-15 | 中国电子科技集团公司第二十八研究所 | Knowledge based indicates the entity relationship prediction technique and forecasting system of study |
| CN109165337B (en) * | 2018-10-17 | 2021-10-15 | 珠海市智图数研信息技术有限公司 | Method and system for establishing bid and ask field association analysis based on knowledge graph |
| CN109165337A (en) * | 2018-10-17 | 2019-01-08 | 珠海市智图数研信息技术有限公司 | A kind of method and system of knowledge based map construction bidding field association analysis |
| CN109614501A (en) * | 2018-12-13 | 2019-04-12 | 浙江工商大学 | A kind of industrial hidden danger standardization report method and system of knowledge based map |
| CN109753571A (en) * | 2019-01-11 | 2019-05-14 | 中山大学 | A kind of scene map lower dimensional space embedding grammar based on secondary theme space projection |
| CN109753571B (en) * | 2019-01-11 | 2022-04-19 | 中山大学 | Scene map low-dimensional space embedding method based on secondary theme space projection |
| CN109922075A (en) * | 2019-03-22 | 2019-06-21 | 中国南方电网有限责任公司 | Network security knowledge map construction method and apparatus, computer equipment |
| CN110275966A (en) * | 2019-07-01 | 2019-09-24 | 科大讯飞(苏州)科技有限公司 | A kind of Knowledge Extraction Method and device |
| CN111392538A (en) * | 2020-03-17 | 2020-07-10 | 浙江新再灵科技股份有限公司 | Elevator comprehensive fault early warning method based on multi-dimensional Internet of things atlas big data |
| CN111680109A (en) * | 2020-04-22 | 2020-09-18 | 北京三快在线科技有限公司 | Knowledge graph representation learning model training method and device and electronic equipment |
| CN111680109B (en) * | 2020-04-22 | 2024-03-29 | 北京三快在线科技有限公司 | Knowledge graph representation learning model training method and device and electronic equipment |
| CN111339321A (en) * | 2020-05-18 | 2020-06-26 | 中国人民解放军国防科技大学 | Triple authenticity detection method and device in knowledge graph |
| CN111339321B (en) * | 2020-05-18 | 2020-08-21 | 中国人民解放军国防科技大学 | Triple authenticity detection method and device in knowledge graph |
| CN112380298A (en) * | 2020-12-03 | 2021-02-19 | 成都航天科工大数据研究院有限公司 | Knowledge graph construction method and system for oil and gas pipeline defect repair |
| CN112784059A (en) * | 2021-01-20 | 2021-05-11 | 和美（深圳）信息技术股份有限公司 | Knowledge graph representation learning method and device, electronic device and storage medium |
| CN113033914A (en) * | 2021-04-16 | 2021-06-25 | 哈尔滨工业大学 | Entity and relation prediction method for machining process knowledge graph |
| CN116705338A (en) * | 2023-08-08 | 2023-09-05 | 中国中医科学院中医药信息研究所 | Traditional Chinese medicine multi-mode knowledge graph reasoning method and device based on rules and paths |
| CN116705338B (en) * | 2023-08-08 | 2023-12-08 | 中国中医科学院中医药信息研究所 | Traditional Chinese medicine multi-mode knowledge graph reasoning method and device based on rules and paths |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106934042B (en) | 2020-05-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106934042B (en) | Knowledge graph representation system and implementation method thereof | |
| CN112308158B (en) | Multi-source field self-adaptive model and method based on partial feature alignment | |
| CN106951499B (en) | A kind of knowledge mapping representation method based on translation model | |
| CN109271522B (en) | Comment emotion classification method and system based on deep hybrid model transfer learning | |
| CN107203787B (en) | Unsupervised regularization matrix decomposition feature selection method | |
| CN111754345B (en) | Bit currency address classification method based on improved random forest | |
| CN107330355B (en) | Deep pedestrian re-identification method based on positive sample balance constraint | |
| CN110110100A (en) | Across the media Hash search methods of discrete supervision decomposed based on Harmonious Matrix | |
| CN109492750B (en) | Zero sample image classification method based on convolutional neural network and factor space | |
| CN113963165B (en) | Small sample image classification method and system based on self-supervision learning | |
| CN107818345A (en) | It is a kind of based on the domain self-adaptive reduced-dimensions method that maximum dependence is kept between data conversion | |
| CN114333064A (en) | Small sample behavior identification method and system based on multidimensional prototype reconstruction reinforcement learning | |
| CN112580502A (en) | SICNN-based low-quality video face recognition method | |
| CN114299362A (en) | Small sample image classification method based on k-means clustering | |
| CN108121962A (en) | Face identification method, device and equipment based on non-negative self-adaptive feature extraction | |
| WO2021128704A1 (en) | Open set classification method based on classification utility | |
| CN110414431B (en) | Face recognition method and system based on elastic context relation loss function | |
| Dai | Imbalanced protein data classification using ensemble FTM-SVM | |
| CN112104602A (en) | Network intrusion detection method based on CNN transfer learning | |
| CN110083724B (en) | Similar image retrieval method, device and system | |
| CN114818963A (en) | Small sample detection algorithm based on cross-image feature fusion | |
| CN113033345B (en) | V2V video face recognition method based on public feature subspace | |
| CN112132841B (en) | Medical image cutting method and device | |
| CN114692809A (en) | Data processing method and device based on neural cluster, storage medium and processor | |
| CN111160398B (en) | Missing label multi-label classification method based on example level and label level association |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |