CN114036330A - Object feature matrix determination method, device, equipment and storage medium - Google Patents

Object feature matrix determination method, device, equipment and storage medium Download PDF

Info

Publication number
CN114036330A
CN114036330A CN202111234766.2A CN202111234766A CN114036330A CN 114036330 A CN114036330 A CN 114036330A CN 202111234766 A CN202111234766 A CN 202111234766A CN 114036330 A CN114036330 A CN 114036330A
Authority
CN
China
Prior art keywords
matrix
heterogeneous
function
attribute
correlation
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
Application number
CN202111234766.2A
Other languages
Chinese (zh)
Inventor
王越辉
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Wingtech Information Technology Co Ltd
Original Assignee
Xian Wingtech Information Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xian Wingtech Information Technology Co Ltd filed Critical Xian Wingtech Information Technology Co Ltd
Priority to CN202111234766.2A priority Critical patent/CN114036330A/en
Priority to PCT/CN2021/140980 priority patent/WO2023065525A1/en
Publication of CN114036330A publication Critical patent/CN114036330A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/55—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/01—Social networking

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Primary Health Care (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • General Health & Medical Sciences (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The disclosure relates to an object feature matrix determination method, apparatus, device, and storage medium. The method comprises the following steps: acquiring heterogeneous characteristic incidence matrixes and corresponding incidence relation indication matrixes of various types of objects in an object data set and multilayer attribute heterogeneous network matrixes of various types of objects; constructing a heterogeneous characteristic correlation function based on the heterogeneous characteristic correlation matrix and the corresponding incidence relation indication matrix; constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix; constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function; calculating low-rank characteristic matrixes of various types of objects based on the minimized objective function; and reconstructing the object feature matrix of each type of object in the object data set by adopting the low-rank feature matrix. By adopting the scheme provided by the embodiment of the disclosure, the information loss caused by homogeneous conversion of the attribute heterogeneous network can be reduced.

Description

Object feature matrix determination method, device, equipment and storage medium
Technical Field
The present disclosure relates to the field of data processing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for determining an object feature matrix.
Background
Efficient classification of objects such as images, biomolecules, social networking site users, etc. relies on establishing an efficient image classifier. In the method for classifying objects by adopting a discriminant model, an effective object classifier is constructed on the premise of extracting an object feature matrix which effectively represents the characteristics of sample objects.
The related technology provides a scheme for determining a low-rank characteristic matrix of an object based on matrix decomposition and constructing the object characteristic matrix of a sample object based on the low-rank characteristic matrix. The scheme of constructing the object feature matrix based on matrix decomposition can ensure the internal structure of the heterogeneous data source. However, due to the imperfection of object data and the limitation of model assumption and experimental design, the existing method still has the problems of information loss caused by homogeneous data conversion and incomplete known heterogeneous associated data.
Disclosure of Invention
In order to solve the technical problem or at least partially solve the technical problem, embodiments of the present disclosure provide an object feature matrix determination method, apparatus, device and storage medium.
In a first aspect, an embodiment of the present disclosure provides an object feature matrix determining method, including:
acquiring heterogeneous characteristic incidence matrixes and corresponding incidence relation indication matrixes of various types of objects in an object data set and multilayer attribute heterogeneous network matrixes of various types of objects;
constructing a heterogeneous characteristic correlation function based on the heterogeneous characteristic correlation matrix and the corresponding incidence relation indication matrix;
constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix;
constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function;
calculating low-rank feature matrices of various types of objects based on the minimized objective function;
and reconstructing an object feature matrix of each type of object in the object data set by using the low-rank feature matrix.
Optionally, the method further comprises:
constructing a first constraint function based on a first weight matrix in the heterogeneous characteristic correlation function;
constructing a second constraint function based on a second weight matrix in the attribute heterogeneous network function;
constructing a minimization objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function, wherein the minimization objective function comprises the following steps:
constructing the minimized objective function based on the heterogeneous feature correlation function, the attribute heterogeneous network function, the first constraint function and the second constraint function.
Optionally, the method further comprises:
further comprising: constructing an unknown noise correlation constraint function;
constructing the minimization objective function based on the heterogeneous feature correlation function, the attribute heterogeneous network function, the first constraint function and the second constraint function, including:
and constructing the minimized objective function based on the heterogeneous characteristic correlation function, the attribute heterogeneous network function, the first constraint function, the second constraint function and the unknown noise correlation constraint function.
Optionally, the determining a low rank feature matrix for each type of object based on the minimization objective function includes:
and solving the minimized objective function by adopting an alternating direction multiplier method, and determining low-rank characteristic matrixes of various types of objects.
The solving the minimized objective function by adopting an alternating direction multiplier method comprises the following steps:
and adopting the alternating direction multiplier method to sequentially and alternately solve the first weight matrix, the second weight matrix, the low-rank characteristic matrix, the heterogeneous correlation network matrix and the base matrix of each multilayer attribute heterogeneous network matrix in the minimized objective function until the preset iteration times or convergence is reached.
Optionally, the alternately solving the first weight matrix, the second weight matrix, the low-rank feature matrix, the heterogeneous association network matrix, and the base matrix of the multi-layer attribute heterogeneous network matrix in the minimized objective function by using the alternate direction multiplier method includes:
setting the first weight matrix, the second weight matrix, the low-rank feature matrix and the heterogeneous correlation network matrix as fixed values, solving a first partial derivative of a base matrix of the multilayer attribute heterogeneous network matrix, and determining the base matrix of the multilayer attribute heterogeneous network matrix when the first partial derivative value is zero;
setting the first weight matrix, the second weight matrix, the low-rank characteristic matrix and the base matrix of the multi-layer attribute heterogeneous network matrix as fixed values, solving a second partial derivative of the heterogeneous correlation network matrix, and determining the heterogeneous correlation network matrix when the second partial derivative value is zero;
setting the first weight matrix, the second weight matrix, the heterogeneous correlation network matrix and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, solving a third partial derivative of the low-rank feature matrix, and determining the low-rank feature matrix when the third partial derivative value is zero;
setting the first weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, solving a fourth partial derivative of the second weight matrix, and determining the second weight matrix when the fourth partial derivative value is zero;
setting the second weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, solving a fifth partial derivative of the first weight matrix, and determining the first weight matrix when the fifth partial derivative value is zero.
Optionally, the object data set is an image data set, and the object in the object data set is an image.
In a second aspect, an embodiment of the present disclosure provides an object feature matrix determining apparatus, including:
the matrix acquisition unit is used for acquiring a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix of various types of objects in the object data set and a multilayer attribute heterogeneous network matrix of various types of objects;
the function construction unit is used for constructing a heterogeneous characteristic incidence function based on the heterogeneous characteristic incidence matrix and the corresponding incidence relation indication matrix; constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix; constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function;
the solving unit is used for calculating low-rank characteristic matrixes of various types of objects based on the minimized objective function;
and the object feature matrix determining unit is used for reconstructing the object feature matrix of each type of object in the object data set by adopting the low-rank feature matrix.
In a third aspect, an embodiment of the present disclosure provides an electronic device, including:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a method as in any one of the first aspects.
In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, where the program is executed by a processor to implement the method according to any one of the first aspect.
Compared with the prior art, the technical scheme provided by the embodiment of the disclosure has the following advantages:
according to the scheme provided by the implementation of the disclosure, a heterogeneous characteristic incidence function is constructed through a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix, so that the heterogeneous incidence and potential heterogeneous incidence are distinguished by utilizing the incidence relation indication matrix. The method comprises the steps of constructing an attribute heterogeneous network function by adopting low-rank feature matrixes corresponding to various types of objects and base matrixes after attribute heterogeneous network decomposition of the various types of objects, constructing a minimized objective function based on a heterogeneous feature correlation function and the attribute heterogeneous network function, cooperatively grading a first weight matrix, a low-rank feature matrix, a heterogeneous correlation network matrix, a second weight matrix in the heterogeneous feature correlation function and the attribute heterogeneous network function and the base matrixes after attribute heterogeneous network decomposition of the various types of objects by utilizing the minimized objective function to obtain the low-rank feature matrixes of the various types of objects, and determining the object feature matrixes of the various types of objects based on the low-rank feature matrixes of the various types of objects. By adopting the scheme provided by the embodiment of the disclosure, the topological structure of the network and the attribute information of the nodes can be fused together, so that the cold start problem caused by the known insufficient association is solved, and the information loss of the attribute heterogeneous network caused by homogeneous conversion is reduced.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.
In order to more clearly illustrate the embodiments or technical solutions in the prior art of the present disclosure, the drawings used in the description of the embodiments or prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a flowchart of an object feature matrix determination method provided in an embodiment of the present disclosure;
FIG. 2 is a flowchart of a method for determining an object feature matrix according to an embodiment of the present disclosure;
FIG. 3 is a flowchart of a method for determining an object feature matrix according to an embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of an object determination apparatus provided in an embodiment of the present disclosure;
fig. 5 is a schematic structural diagram of an electronic device provided in an embodiment of the present disclosure.
Detailed Description
In order that the above objects, features and advantages of the present disclosure may be more clearly understood, aspects of the present disclosure will be further described below. It should be noted that the embodiments and features of the embodiments of the present disclosure may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure, but the present disclosure may be practiced in other ways than those described herein; it is to be understood that the embodiments disclosed in the specification are only a few embodiments of the present disclosure, and not all embodiments.
The embodiment of the disclosure provides an object feature matrix determining method, which is used for solving the problems of information loss caused by homogeneous data conversion and incomplete known heterogeneous data association in the fusion (namely matrix decomposition) process of the existing method for determining object features based on matrix decomposition, and then more accurately characterizing the features of objects in an object data set. In particular embodiments, the object in the embodiments of the present disclosure may be an image, a biomolecule, a social networking site user, or the like.
The object feature matrix determination method of the present disclosure is executed by an electronic device or an application program in the electronic device. The electronic device may be a tablet computer, a mobile phone, a notebook computer, a server, etc., and the present disclosure does not set any limit to the specific type of the electronic device. The present disclosure is not limited as to the type of operating system of the electronic device. For example, an Android system, a Linux system, a Windows system, an iOS system, etc.
Fig. 1 is a flowchart of an object feature matrix determination method provided in an embodiment of the present disclosure. As shown in fig. 1, the object feature matrix determination method provided by the embodiment of the present disclosure includes steps S101 to S106.
Step S101: and acquiring a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix of each type of object in the object data set, and a multilayer attribute heterogeneous network matrix of each type of object.
In the embodiment of the present disclosure, the electronic device may process raw data of various types of objects in the object data set, and obtain a heterogeneous characteristic association matrix, a corresponding association relation indication matrix, and a multi-layer attribute heterogeneous network matrix of the various types of objects, or may obtain the above matrix by reading data pre-stored in a storage.
The disclosed embodiments adopt
Figure BDA0003317325330000061
N representing the ith type of objectiA heterogeneous feature correlation matrix between the individual samples,by means of HijIs represented by the formulaijCorresponding and having dimension of RijThe same incidence relation indicates a matrix. Wherein the incidence relation indication matrix HijFor distinguishing observed and unobserved correlations in a consistent feature correlation matrix. In particular, if R isij(s,t)>0, then Hij1, otherwise Hij＝0。
The disclosed embodiments adopt
Figure BDA0003317325330000062
Multiple attribute heterogeneous network matrix representing the ith type of object, where tiIndicates that the ith type object collects t togetheriHeterogeneous network of attributes of species origin, ditIndicates that the t-th attribute heterogeneous network has ditAnd (6) a kind of attribute.
Step S102: and constructing a heterogeneous characteristic correlation function based on the heterogeneous characteristic correlation matrix and the corresponding incidence relation indication matrix.
The heterogeneous characteristic correlation function comprises a first weight matrix, a low-rank characteristic matrix corresponding to each type of object and a heterogeneous correlation network matrix among each type of object besides the heterogeneous characteristic correlation matrix and the corresponding correlation management indication matrix.
In one embodiment of the present disclosure, the heterogeneous characteristic correlation function is represented by equation 1.1.
Figure BDA0003317325330000071
Wherein: e represents a Hadamard product;
Figure BDA0003317325330000072
representing a low-rank characteristic matrix of heterogeneous associated network guidance between ith or jth type objects, wherein two low-rank identification matrices respectively represent ki and k in compressionjN for describing i type object in real number form by dimensional spaceiAttribute information of individual object and n of j-th type objectjProperties of individual objectsInformation; k is a radical ofiAnd kjRespectively representing the dimensionality of the ith or jth low-rank feature matrix;
Figure BDA0003317325330000073
shows a heterogeneous correlation network matrix, compared to RijThe scale is much smaller;
Figure BDA00033173253300000711
presentation pair
Figure BDA0003317325330000074
A first weight matrix assigned to the heterogeneous correlation network
Figure BDA0003317325330000075
Figure BDA0003317325330000076
Can be approximated as reconstruction loss and used to distinguish observed associations from unobserved associations and to make the observed associations R after reconstructionijIs maintained.
Step S103: and constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix.
In the embodiment of the present disclosure, the attribute heterogeneous network function includes, in addition to the aforementioned multiple attribute heterogeneous network matrices, a second weight matrix, low-rank feature matrices corresponding to various types of objects, and a base matrix after attribute heterogeneous network decomposition of various types of objects.
In one embodiment of the present disclosure, the attribute heterogeneous network function is represented by equation 1.2.
Figure BDA0003317325330000077
Wherein,
Figure BDA0003317325330000078
representing second weight moments assigned to the attribute heterogeneous network matrix of the plurality of objectsArray, for XitIf t is>maxiti,
Figure BDA0003317325330000079
GiA low-rank feature matrix representing attribute heterogeneous network guidance of the ith type of object;
Figure BDA00033173253300000710
and representing the decomposed base matrix of the attribute heterogeneous network of the ith type object.
In the embodiment of the disclosure, the attribute heterogeneous network function is adopted to directly decompose the attribute heterogeneous data, so that information loss caused by homogeneous conversion is avoided.
Step S104: and constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function.
In the embodiment of the present disclosure, the minimization objective function is constructed based on the heterogeneous feature correlation function and the attribute heterogeneous network function, and the heterogeneous feature correlation function and the attribute heterogeneous network function may be added.
In one embodiment of the present disclosure, after adding the heterogeneous characteristic correlation function and the attribute heterogeneous network function, the obtained minimized objective function is expressed by equation 1.3
Figure BDA0003317325330000081
s.t.ωr≥0,ωh≥0,∑vec(ωr)＝1,∑vec(ωh)＝1 (1.3)
Step S105: and calculating low-rank feature matrixes of various types of objects based on the minimized objective function.
In the embodiment of the disclosure, the calculation of the low-rank feature matrices of the various types of objects based on the minimization target function is to perform collaborative decomposition on a first weight matrix in the minimization target function, a low-rank feature matrix corresponding to the various types of objects, a heterogeneous association network matrix among the various types of objects, a second weight matrix, and a base matrix after attribute heterogeneous network decomposition of the various types of objects, and obtain a low-rank feature matrix G of the various types of objects in the collaborative decomposition process.
In the disclosed embodiment, the objective function is minimized in G, S, U, ωr,ωhThe upper surface is non-convex and therefore can be solved optimally by means of an Alternating Direction Multiplier Method (ADMM) which is used to approximate the three-factor matrix decomposition. Specifically, G, S, U, ω can be representedr,ωhThe other parameter is optimized at the same time, and iteration is repeated until all the parameters are solved. Specifically, the method calculates G, S, U and omega by adopting an alternative direction multiplier methodr,ωhAs will be described hereinafter.
Step S106: and reconstructing the object feature matrix of each type of object in the object data set by adopting the low-rank feature matrix.
After determining the low-rank feature matrices for various types of objects, the low-rank feature matrices may be employed to reconstruct the object feature matrices. In some embodiments of the present disclosure, after determining the low-rank feature matrices of the various types of objects, the low-rank feature matrices may be directly used as object feature matrices of the corresponding types of objects.
According to the object feature matrix determining method, a heterogeneous feature correlation function is established through a heterogeneous feature correlation matrix and a corresponding correlation indication matrix, so that heterogeneous correlation and potential heterogeneous correlation are distinguished by using the correlation indication matrix. The method comprises the steps of constructing an attribute heterogeneous network function by adopting low-rank feature matrixes corresponding to various types of objects and base matrixes after attribute heterogeneous network decomposition of the various types of objects, constructing a minimized objective function based on a heterogeneous feature correlation function and the attribute heterogeneous network function, cooperatively grading a first weight matrix, a low-rank feature matrix, a heterogeneous correlation network matrix, a second weight matrix in the heterogeneous feature correlation function and the attribute heterogeneous network function and the base matrixes after attribute heterogeneous network decomposition of the various types of objects by utilizing the minimized objective function to obtain the low-rank feature matrixes of the various types of objects, and determining the object feature matrixes of the various types of objects based on the low-rank feature matrixes of the various types of objects.
By adopting the method provided by the embodiment of the disclosure, the topological structure of the network and the attribute information of the nodes can be fused together, thereby making up the problem of cold start caused by known insufficient association and further reducing the information loss of the attribute heterogeneous network caused by homogeneous conversion.
Fig. 2 is a flowchart of a method for determining an object feature matrix according to an embodiment of the present disclosure. As shown in fig. 2, the method provided by the embodiment of the present disclosure includes steps S201 to S207.
Step S201: and acquiring a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix of each type of object in the object data set, and a multilayer attribute heterogeneous network matrix of each type of object.
Step S202: and constructing a heterogeneous characteristic correlation function based on the heterogeneous characteristic correlation matrix and the corresponding incidence relation indication matrix.
Step S203: and constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix.
The foregoing steps S201 to S203 are the same as the steps S101 to S103 in the foregoing embodiment, and reference may be made to the foregoing embodiment specifically, and the description is not repeated here.
Step S204: constructing a first constraint function based on a first weight matrix in the heterogeneous characteristic correlation function; and constructing a second constraint function based on a second weight matrix in the attribute heterogeneous network function.
Step S205: and constructing a minimized objective function based on the heterogeneous characteristic correlation function, the attribute heterogeneous network function, the first constraint function and the second constraint function.
When solving the minimized objective function shown in equation 1.3, R is the ratio of the minimum to the maximumijWith minimum approximate loss
Figure BDA0003317325330000101
When R isijWill be weighted as
Figure BDA0003317325330000102
At which time the other heterogeneous incidence matrices will all be ignored. Similarly, whenXitWith minimum approximate loss
Figure BDA0003317325330000103
At this time, the tendency is that
Figure BDA0003317325330000104
Is assigned to Xit. That is, the contributions of all other homogeneous matrices are ignored.
At the same time, a great deal of research has demonstrated that different data sources can provide complementary information with respect to each other. Thus, using only a single heterogeneous correlation matrix and a single homogeneous correlation matrix may not give a reliable prediction.
In order to compensate for the foregoing disadvantage of weight distribution, in the embodiment of the present disclosure, a first constraint function is first constructed based on a first weight matrix in the heterogeneous characteristic correlation function, and a second constraint function is constructed based on a second weight matrix in the attribute heterogeneous network function.
In the embodiment of the disclosure, the first constraint function and the second constraint function are both regular terms based on the norm of l 2. The first constraint function is
Figure BDA0003317325330000105
The second constraint function is
Figure BDA0003317325330000106
The aforementioned vec (ω)r) Is to berThe line-stacked and spliced vector of vec (ω)h) Is to behThe rows of (a) are stacked up with the spliced vectors. Alpha is alpha>0,β>0 is used to control vec (ω)r) And vec (ω)h) Of the system. Meanwhile, the alpha and the beta can also help to selectively integrate different heterogeneous associated data sources and attribute heterogeneous data sources.
In the case of adding the first constraint function and the second constraint function, the minimized objective function obtained in the embodiment of the present disclosure is expressed by equation 1.4.
Figure BDA0003317325330000107
Step S206: and calculating low-rank feature matrixes of various types of objects based on the minimized objective function.
Step S207: and reconstructing an object feature matrix of each type of object in the object data set by using the low-rank feature matrix.
The foregoing steps S206-S207 are the same as the steps S105-S106 of the foregoing embodiment, and reference may be made to the foregoing embodiment specifically, and the description is not repeated here.
The object feature matrix determining method provided by the embodiment of the disclosure constructs a first constraint function based on a first weight matrix in a heterogeneous feature correlation function, constructs a second constraint function based on a second weight matrix in an attribute heterogeneous network function, constructs a minimized objective function through the first constraint function and the second constraint function, the heterogeneous feature correlation function and the attribute heterogeneous network function, and calculates and obtains features of each object based on the minimized objective function, thereby overcoming the defect that reliable prediction may not be given only by using a single heterogeneous correlation matrix and a single homogeneous correlation matrix.
Fig. 3 is a flowchart of a method for determining an object according to an embodiment of the present disclosure. As shown in fig. 3, the method provided by the embodiment of the present disclosure includes steps S301 to S308.
Step S301: and acquiring a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix of each type of object in the object data set, and a multilayer attribute heterogeneous network matrix of each type of object.
Step S302: and constructing a heterogeneous characteristic correlation function based on the heterogeneous characteristic correlation matrix and the corresponding incidence relation indication matrix.
Step S303: and constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix.
Step S304: constructing a first constraint function based on a first weight matrix in the heterogeneous characteristic correlation function; and constructing a second constraint function based on a second weight matrix in the attribute heterogeneous network function.
The foregoing steps S301 to S304 are the same as the steps S201 to S204 in the foregoing embodiment, and reference may be made to the foregoing embodiment specifically, and the description is not repeated here.
Step S305: and constructing an unknown noise correlation constraint function.
In a specific embodiment, the incidence relation indication matrix HijObserved and unobserved correlations are distinguished, but only observed correlations are constrained, so that a great deal of unobserved correlations are likely to be left at random to some extent unconstrained, and a great deal of noise appears in the target correlation matrix.
In order to avoid the foregoing problem, in the embodiments of the present disclosure, an unknown noise correlation constraint function is also constructed, and a minimization objective function is constructed based on the unknown noise correlation constraint function.
In the disclosed embodiment, the unknown noise correlation constraint function is
Figure BDA0003317325330000121
Wherein
Figure BDA0003317325330000122
Is a constraint on the unknown noise correlation and gamma is used to control the complexity of this term.
It should be noted that the foregoing steps S304 and S305 may be performed sequentially or in parallel.
Step S306: and constructing a minimized objective function based on the heterogeneous characteristic correlation function, the attribute heterogeneous network function, the first constraint function, the second constraint function and the unknown noise correlation constraint function.
After the unknown noise association constraint function is determined, the minimized objective function constructed in the embodiment of the disclosure is
Figure BDA0003317325330000123
Step S307: and calculating low-rank feature matrixes of various types of objects based on the minimized objective function.
Step S308: and reconstructing the object feature matrix of each type of object in the object data set by adopting the low-rank feature matrix.
The foregoing steps S307 to S308 are the same as the steps S206 to S207 of the foregoing embodiment, and reference may be made to the foregoing embodiment specifically, and the description is not repeated here.
In the implementation process of the present disclosure, the minimization objective function may be solved by using an alternating direction multiplier method, so as to determine low-rank feature matrices of various types of objects. Specifically, the low-rank characteristic matrix of each type of object is calculated by adopting an alternating direction multiplier method, and a first weight matrix, a second weight matrix, the low-rank characteristic matrix, the heterogeneous correlation network matrix and the base matrix of each multilayer attribute heterogeneous network matrix in the minimized objective function are sequentially and alternately solved by adopting the alternating direction multiplier method until the preset iteration times or convergence is reached.
In the embodiment of the present disclosure, solving the minimized objective function by using the alternating direction multiplier method may include steps S401 to S405. The alternative direction multiplier algorithm employed in the embodiments of the present disclosure is analyzed below by taking the solution of equation 1.5 as an example.
Before calculating equation 1.5, first introduce constraint GiLagrange multiplier of 0 or more
Figure BDA0003317325330000131
Equation 1.5 may be equivalent to equation 1.6.
Figure BDA0003317325330000132
Steps S401-S405 may then be performed using an alternating direction multiplier algorithm based on equation 1.6.
Step S401: setting the first weight matrix, the second weight matrix, the low-rank characteristic matrix and the heterogeneous correlation network matrix as fixed values, solving a first partial derivative of the base matrix of the multilayer attribute heterogeneous network matrix, and determining the base matrix of the multilayer attribute heterogeneous network matrix when the first partial derivative value is zero.
In particular, assume G, S, ωr,ωhIt is known that U can be optimizeditThus, equation 1.6 is related to UitPartial derivatives of (a):
Figure BDA0003317325330000133
for the
Figure BDA0003317325330000134
So that
Figure BDA0003317325330000135
Can obtain
Figure BDA0003317325330000136
I.e., the base matrix of the multi-layer attribute heterogeneous network matrix is
Figure BDA0003317325330000137
Step S402: setting the base matrixes of the first weight matrix, the second weight matrix, the low-rank characteristic matrix and the multi-layer attribute heterogeneous network matrix as fixed values, solving a second partial derivative of the heterogeneous correlation network matrix, and determining the heterogeneous correlation network matrix when the second partial derivative value is zero.
In particular, assume G, U, ωr,ωhIt is known that S can be optimizedijThus, equation 1.6 is related to SijThe partial derivative of (c) yields:
Figure BDA0003317325330000141
for the
Figure BDA0003317325330000142
So that
Figure BDA0003317325330000143
It is possible to obtain:
Figure BDA0003317325330000144
i.e. heterogeneous associative network matrix of
Figure BDA0003317325330000145
Step S403: setting the base matrixes of the first weight matrix, the second weight matrix, the heterogeneous correlation network matrix and the multilayer attribute heterogeneous network matrix as fixed values, solving a third partial derivative of the low-rank characteristic matrix, and determining the low-rank characteristic matrix when the third partial derivative value is zero.
In particular, assume S, U, ωr,ωhIt is known that formula 1.6 can be related to GiPartial derivatives of (a):
Figure BDA0003317325330000146
polynomial factor lambdaiCan pass through
Figure BDA0003317325330000147
Obtained under the conditions of Karush-Kuhn-Tucker (KKT):
Figure BDA0003317325330000148
where e represents the Hadamard product, equation 1.10 is a fixed point equation and the solution must satisfy the convergence condition, so that:
Figure BDA0003317325330000151
for the
Figure BDA0003317325330000152
Figure BDA0003317325330000153
For t ═ 1,2, K, maxiti：
Figure BDA0003317325330000154
The signs of positive and negative values in equations 1.11, 1.12, and 1.13 may be defined as
Figure BDA0003317325330000155
And
Figure BDA0003317325330000156
thus, the low rank feature matrix G may be updated as:
Figure BDA0003317325330000157
step S404: setting the base matrixes of the first weight matrix, the low-rank characteristic matrix, the heterogeneous correlation network matrix and the multilayer attribute heterogeneous network matrix as fixed values, solving a fourth partial derivative of the second weight matrix, and determining the second weight matrix when the fourth partial derivative value is zero.
In particular, in the pair omegahWhen calculating the partial derivatives, the 1 st, 3 rd and 5 th parts on the right side of the formula 1.6 and omegahIrrelevant and therefore negligible. Further, it is possible to obtain:
Figure BDA0003317325330000158
order to
Figure BDA0003317325330000159
Representing an Attribute heterogeneous network XitThe aforementioned formula can be simplified to:
Figure BDA00033173253300001510
the solution to equation 1.15 can be viewed as relating to vec (ω)h) The quadratic programming problem of (2) can be solved by introducing lagrangian multipliers based on an algorithm of Selective Non-Matrix Factorization (SNMF).
Step S405: setting the base matrixes of the second weight matrix, the low-rank characteristic matrix, the heterogeneous correlation network matrix and the multilayer attribute heterogeneous network matrix as fixed values, solving a fifth partial derivative of the first weight matrix, and determining the first weight matrix when the fifth partial derivative value is zero.
Specifically, when it is against ωrWhen calculating the partial derivatives, the 2 nd, 4 th and 5 th parts on the right side of the formula 1.6 and omegarIrrelevant and therefore negligible. Further, it is possible to obtain:
Figure BDA0003317325330000161
order to
Figure BDA0003317325330000162
Representing a heterogeneous association RijEquation 1.16 can be simplified to:
Figure BDA0003317325330000163
equation 1.17 can be viewed as relating to vec (ω)r) The quadratic programming problem can also be solved by introducing a Lagrange multiplier based on the algorithm of selective matrix decomposition. It should be noted that the foregoing steps S401-S405 are not limited sequentially in the specific embodiment.
On the basis of the above embodiments, the embodiments of the present disclosure provide an object determination apparatus. Fig. 4 is a schematic structural diagram of an object determination apparatus according to an embodiment of the present disclosure. As shown in fig. 4, the object determining apparatus provided in the embodiment of the present disclosure includes a matrix obtaining unit 401, a function constructing unit 402, a solving unit 403, and an object feature matrix determining unit 404.
The matrix obtaining unit 401 is configured to obtain a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix of each type of object in the object data set, and a multi-layer attribute heterogeneous network matrix of each type of object;
the function building unit 402 is configured to build a heterogeneous feature correlation function based on the heterogeneous feature correlation matrix and the corresponding incidence relation indication matrix; constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix; constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function;
the solving unit 403 is configured to calculate low-rank feature matrices of various types of objects based on a minimized objective function;
the object feature matrix determination unit 404 is configured to reconstruct an object feature matrix of each type of object in the object data set using the low-rank feature matrix.
The object feature matrix determining apparatus provided in the embodiment of the present disclosure constructs a heterogeneous feature correlation function through the heterogeneous feature correlation matrix and the corresponding correlation indication matrix, so as to distinguish heterogeneous correlations and potential heterogeneous correlations by using the correlation indication matrix. The method comprises the steps of constructing an attribute heterogeneous network function by adopting low-rank feature matrixes corresponding to various types of objects and base matrixes after attribute heterogeneous network decomposition of the various types of objects, constructing a minimized objective function based on a heterogeneous feature correlation function and the attribute heterogeneous network function, cooperatively grading a first weight matrix, a low-rank feature matrix, a heterogeneous correlation network matrix, a second weight matrix in the heterogeneous feature correlation function and the attribute heterogeneous network function and the base matrixes after attribute heterogeneous network decomposition of the various types of objects by utilizing the minimized objective function to obtain the low-rank feature matrixes of the various types of objects, and determining the object feature matrixes of the various types of objects based on the low-rank feature matrixes of the various types of objects.
By adopting the device provided by the embodiment of the disclosure, the topological structure of the network and the attribute information of the nodes can be fused together, thereby making up the problem of cold start caused by the known insufficient association and further reducing the information loss of the attribute heterogeneous network caused by homogeneous conversion.
In some embodiments of the present disclosure, the function construction unit 402 is further configured to construct a first constraint function based on a first weight matrix in the heterogeneous feature correlation function, and to construct a second constraint function based on a second weight matrix in the attribute heterogeneous network function. Correspondingly, the function construction unit 402 constructs a minimization objective function based on the heterogeneous feature correlation function, the attribute heterogeneous network function, the first constraint function, and the second constraint function.
In some embodiments of the present disclosure, the function construction unit 402 is further configured to construct an unknown noise correlation constraint function. Correspondingly, the function construction unit 402 constructs a minimized objective function based on the heterogeneous feature correlation function, the attribute heterogeneous network function, the first constraint function, the second constraint function, and the unknown noise correlation constraint function.
In some embodiments of the present disclosure, the solving unit 403 solves the minimized objective function by using an alternating direction multiplier method, and determines a low-rank feature matrix of each type of object.
Specifically, in some embodiments of the present disclosure, the solving unit 403 sequentially and alternately solves the first weight matrix, the second weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix, and the base matrix of each multi-layer attribute heterogeneous network matrix in the minimized objective function by using an alternating direction multiplier method until reaching a preset iteration number or converging.
In practical implementation, the solving unit 403 calculates a first partial derivative for the base matrix of the multi-layer attribute heterogeneous network matrix by setting the first weight matrix, the second weight matrix, the low-rank feature matrix and the heterogeneous associated network matrix as fixed values, and determines the base matrix of the multi-layer attribute heterogeneous network matrix when the first partial derivative value is zero;
in practical implementation, the solving unit 403 finds a second partial derivative for the heterogeneous associated network matrix by setting the first weight matrix, the second weight matrix, the low-rank feature matrix and the base matrix of the multi-layer attribute heterogeneous network matrix as fixed values, and determines the heterogeneous associated network matrix when the second partial derivative value is zero;
in practical implementation, the solving unit 403 finds a third partial derivative for the low-rank feature matrix by setting the first weight matrix, the second weight matrix, the heterogeneous correlation network matrix, and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, and determines the low-rank feature matrix when the third partial derivative value is zero;
in practical implementation, the solving unit 403 finds a fourth partial derivative for the second weight matrix by setting the first weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix, and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, and determines the second weight matrix when the fourth partial derivative value is zero;
in practical implementation, the solving unit 403 finds the fifth partial derivative of the first weight matrix by setting the second weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix, and the base matrix of the multi-layer attribute heterogeneous network matrix as fixed values, and determines the first weight matrix when the fifth partial derivative value is zero.
In an embodiment of the disclosure, the object data set may be an image data set, and the object in the object data set is an image.
The device provided by the embodiment of the invention can execute the method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Fig. 5 is a schematic structural diagram of an electronic device provided in an embodiment of the present disclosure, as shown in fig. 5, the electronic device includes a processor 501, a memory 502, an input device 503, and an output device 504; the number of the processors 501 in the electronic device may be one or more, and one processor 501 is taken as an example in fig. 5; the processor 501, the memory 502, the input device 503 and the output device 504 in the electronic apparatus may be connected by a bus or other means, and the connection by the bus is exemplified in fig. 5.
The memory 502 is a computer-readable storage medium that can be used to store software programs, computer-executable programs, and modules, such as program instructions/modules corresponding to the methods in the embodiments of the present invention. The processor 501 executes various functional applications and data processing of the electronic device by executing software programs, instructions and modules stored in the memory 502, so as to implement the method provided by the embodiment of the invention.
The memory 502 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required for at least one function; the storage data area may store data created according to the use of the terminal, and the like. Further, the memory 502 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid state storage device. In some examples, memory 502 may further include memory located remotely from processor 501, which may be connected to a computer device through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The input device 503 may be used to receive input numeric or character information and generate key signal inputs related to user settings and function control of the electronic apparatus, and may include a keyboard, a mouse, etc., and the output device 504 may include a display device such as a display screen.
The disclosed embodiments also provide a storage medium containing computer-executable instructions, which when executed by a computer processor, are used to implement the methods provided by the embodiments of the present invention.
Of course, the storage medium provided by the embodiment of the present invention contains computer-executable instructions, and the computer-executable instructions are not limited to the above method operations, and may also perform related operations in the method provided by any embodiment of the present invention.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present disclosure, which enable those skilled in the art to understand or practice the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the disclosure. Thus, the present disclosure is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. An object feature matrix determination method, comprising:
acquiring heterogeneous characteristic incidence matrixes and corresponding incidence relation indication matrixes of various types of objects in an object data set and multilayer attribute heterogeneous network matrixes of various types of objects;
constructing a heterogeneous characteristic correlation function based on the heterogeneous characteristic correlation matrix and the corresponding incidence relation indication matrix;
constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix;
constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function;
calculating low-rank feature matrices of various types of objects based on the minimized objective function;
and reconstructing an object feature matrix of each type of object in the object data set by using the low-rank feature matrix.
2. The method of claim 1, further comprising:
constructing a first constraint function based on a first weight matrix in the heterogeneous characteristic correlation function;
constructing a second constraint function based on a second weight matrix in the attribute heterogeneous network function;
constructing a minimization objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function, wherein the minimization objective function comprises the following steps:
constructing the minimized objective function based on the heterogeneous feature correlation function, the attribute heterogeneous network function, the first constraint function and the second constraint function.
3. The method of claim 2, further comprising: constructing an unknown noise correlation constraint function;
constructing the minimization objective function based on the heterogeneous feature correlation function, the attribute heterogeneous network function, the first constraint function and the second constraint function, including:
and constructing the minimized objective function based on the heterogeneous characteristic correlation function, the attribute heterogeneous network function, the first constraint function, the second constraint function and the unknown noise correlation constraint function.
4. The method according to any of claims 1-3, wherein said determining a low rank feature matrix for each type of object based on said minimization objective function comprises:
and solving the minimized objective function by adopting an alternating direction multiplier method, and determining low-rank characteristic matrixes of various types of objects.
5. The method of claim 4, wherein solving the minimized objective function using an alternating direction multiplier method comprises:
and adopting the alternating direction multiplier method to sequentially and alternately solve the first weight matrix, the second weight matrix, the low-rank characteristic matrix, the heterogeneous correlation network matrix and the base matrix of each multilayer attribute heterogeneous network matrix in the minimized objective function until the preset iteration times or convergence is reached.
6. The method of claim 5, wherein the alternating direction multiplier method is adopted to sequentially and alternately solve the first weight matrix, the second weight matrix, the low rank feature matrix, the heterogeneous correlation network matrix and the base matrix of the multi-layer property heterogeneous network matrix in the minimized objective function, and comprises:
setting the first weight matrix, the second weight matrix, the low-rank feature matrix and the heterogeneous correlation network matrix as fixed values, solving a first partial derivative of a base matrix of the multilayer attribute heterogeneous network matrix, and determining the base matrix of the multilayer attribute heterogeneous network matrix when the first partial derivative value is zero;
setting the first weight matrix, the second weight matrix, the low-rank characteristic matrix and the base matrix of the multi-layer attribute heterogeneous network matrix as fixed values, solving a second partial derivative of the heterogeneous correlation network matrix, and determining the heterogeneous correlation network matrix when the second partial derivative value is zero;
setting the first weight matrix, the second weight matrix, the heterogeneous correlation network matrix and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, solving a third partial derivative of the low-rank feature matrix, and determining the low-rank feature matrix when the third partial derivative value is zero;
setting the first weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, solving a fourth partial derivative of the second weight matrix, and determining the second weight matrix when the fourth partial derivative value is zero;
setting the second weight matrix, the low-rank feature matrix, the heterogeneous correlation network matrix and the base matrix of the multilayer attribute heterogeneous network matrix as fixed values, solving a fifth partial derivative of the first weight matrix, and determining the first weight matrix when the fifth partial derivative value is zero.
7. The method according to any of claims 1-3, wherein the object data set is an image data set and the object in the object data set is an image.
8. An object feature matrix determination apparatus, comprising:
the matrix acquisition unit is used for acquiring a heterogeneous characteristic incidence matrix and a corresponding incidence relation indication matrix of various types of objects in the object data set and a multilayer attribute heterogeneous network matrix of various types of objects;
the function construction unit is used for constructing a heterogeneous characteristic incidence function based on the heterogeneous characteristic incidence matrix and the corresponding incidence relation indication matrix; constructing an attribute heterogeneous network function based on the multilayer attribute heterogeneous network matrix; constructing a minimized objective function based on the heterogeneous characteristic correlation function and the attribute heterogeneous network function;
the solving unit is used for calculating low-rank characteristic matrixes of various types of objects based on the minimized objective function;
and the object feature matrix determining unit is used for reconstructing the object feature matrix of each type of object in the object data set by adopting the low-rank feature matrix.
9. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement a method as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which program, when being executed by a processor, carries out the method of any one of claims 1 to 7.
CN202111234766.2A 2021-10-22 2021-10-22 Object feature matrix determination method, device, equipment and storage medium Pending CN114036330A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202111234766.2A CN114036330A (en) 2021-10-22 2021-10-22 Object feature matrix determination method, device, equipment and storage medium
PCT/CN2021/140980 WO2023065525A1 (en) 2021-10-22 2021-12-23 Object feature matrix determination method and apparatus, device, and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202111234766.2A CN114036330A (en) 2021-10-22 2021-10-22 Object feature matrix determination method, device, equipment and storage medium

Publications (1)

Publication Number Publication Date
CN114036330A true CN114036330A (en) 2022-02-11

Family

ID=80135208

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202111234766.2A Pending CN114036330A (en) 2021-10-22 2021-10-22 Object feature matrix determination method, device, equipment and storage medium

Country Status (2)

Country Link
CN (1) CN114036330A (en)
WO (1) WO2023065525A1 (en)

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8935308B2 (en) * 2012-01-20 2015-01-13 Mitsubishi Electric Research Laboratories, Inc. Method for recovering low-rank matrices and subspaces from data in high-dimensional matrices
US10229092B2 (en) * 2017-08-14 2019-03-12 City University Of Hong Kong Systems and methods for robust low-rank matrix approximation
CN110222213B (en) * 2019-05-28 2021-07-16 天津大学 Image classification method based on heterogeneous tensor decomposition
CN111310813A (en) * 2020-02-07 2020-06-19 广东工业大学 Subspace clustering method and device for potential low-rank representation

Also Published As

Publication number Publication date
WO2023065525A1 (en) 2023-04-27

Similar Documents

Publication Publication Date Title
US11854671B2 (en) Method and apparatus for identifying heterogeneous graph and property of molecular space structure and computer device
CN112199189B (en) Adaptation of deep learning model to resource-constrained edge devices
CN109934249A (en) Data processing method, device, medium and calculating equipment
CN111737535B (en) Network characterization learning method based on element structure and graph neural network
CN111177473B (en) Personnel relationship analysis method, device and readable storage medium
CN113780584B (en) Label prediction method, label prediction device, and storage medium
WO2023045149A1 (en) Image fusion method and apparatus, electronic device, and storage medium
CN109992676B (en) Cross-media resource retrieval method and retrieval system
CN110837577A (en) Video recommendation method, device, equipment and storage medium
CN112231570A (en) Recommendation system trust attack detection method, device, equipment and storage medium
CN112561031A (en) Model searching method and device based on artificial intelligence and electronic equipment
WO2023045137A1 (en) Data fusion method and apparatus, electronic device, and computer readable storage medium
CN110990624A (en) Video recommendation method, device, equipment and storage medium
CN114580794A (en) Data processing method, apparatus, program product, computer device and medium
CN110232151B (en) Construction method of QoS (quality of service) prediction model for mixed probability distribution detection
CN112270950A (en) Fusion network drug target relation prediction method based on network enhancement and graph regularization
WO2024179409A1 (en) Three-dimensional industrial anomaly detection method and apparatus, storage medium, and electronic device
WO2024179409A9 (en) Three-dimensional industrial anomaly detection method and apparatus, storage medium, and electronic device
CN117992805A (en) Zero sample cross-modal retrieval method and system based on tensor product graph fusion diffusion
JP7508572B2 (en) Predictive Maintenance Convolutional Neural Networks
CN111814044B (en) Recommendation method, recommendation device, terminal equipment and storage medium
CN116956996A (en) Data processing method, device, medium and equipment
CN114036330A (en) Object feature matrix determination method, device, equipment and storage medium
CN114663401A (en) Defect prediction method and system for small sample knowledge transfer learning
CN113569890A (en) Data processing method and related device

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