CN107391619A - For the adaptive hash method and device of imperfect isomeric data - Google Patents
For the adaptive hash method and device of imperfect isomeric data Download PDFInfo
- Publication number
- CN107391619A CN107391619A CN201710544048.2A CN201710544048A CN107391619A CN 107391619 A CN107391619 A CN 107391619A CN 201710544048 A CN201710544048 A CN 201710544048A CN 107391619 A CN107391619 A CN 107391619A
- Authority
- CN
- China
- Prior art keywords
- adaptive
- hash
- unified
- mode
- isomeric data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000005295 random walk Methods 0.000 claims abstract description 9
- 238000010276 construction Methods 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification 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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/901—Indexing; Data structures therefor; Storage structures
- G06F16/9014—Indexing; Data structures therefor; Storage structures hash tables
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of adaptive hash method and device for imperfect isomeric data, wherein, method includes：When isomeric data sample lacks, bipartite graph progress Markov random walk is built respectively to each mode of multiple mode and probability transfer calculates, and estimates the adjacency information for being related to lacking sample；The adjacency information that each mode for merging multiple mode is practised by adaptive graphics obtains unified adjacency information, and obtains unified Hash by anchor salted hash Salted according to unified adjacency information and encode.This method can effectively lift the information retrieval performance of salted hash Salted, and obtaining unified Hash by study encodes, and so as to the retrieval effectiveness of lifting system, reduces unnecessary information misguidance, reduce the inaccuracy, not scientific for manually setting and bringing.
Description
Technical field
The present invention relates to technical field of information retrieval, more particularly to a kind of adaptive Hash for imperfect isomeric data
Method and device.
Background technology
At present, it is a kind of common application to carry out information retrieval to isomeric data using salted hash Salted, in correlation technique, isomery
Each mode of data is all complete, does not lack any sample.But actually each mode have missing sample can
Energy property, therefore, the salted hash Salted in correlation technique are difficult to obtain preferably retrieval performance on incomplete isomeric data, need
Improve.
The content of the invention
It is contemplated that at least solves one of technical problem in correlation technique to a certain extent.
Therefore, it is an object of the present invention to propose a kind of adaptive hash method for imperfect isomeric data,
This method can lifting system retrieval effectiveness, reduce unnecessary information misguidance, reduce the inaccuracy, no for manually setting and bringing
It is scientific.
It is another object of the present invention to propose a kind of adaptive Hash device for imperfect isomeric data.
To reach above-mentioned purpose, one aspect of the present invention embodiment proposes a kind of for the adaptive of imperfect isomeric data
Hash method, comprise the following steps：When isomeric data sample lacks, bipartite graph is built respectively to each mode of multiple mode
Carry out Markov random walk and probability transfer calculates, and estimate the adjacency information for being related to lacking sample；By adaptive
Answer graphics to practise the adjacency information for each mode for merging the multiple mode and obtain unified adjacency information, and according to the unification
Adjacency information unified Hash obtained by anchor salted hash Salted encoded.
The adaptive hash method for imperfect isomeric data of the embodiment of the present invention, in a unified vague generalization frame
In frame, the retrieval effectiveness of system is improved by the Hash coding techniques of linear session, wherein, it is related to lacking sample in estimation
Adjacency information when, it is only necessary to utilize existing data, it is not necessary to auxiliary information is additionally introduced, so as to reduce unnecessary letter
Breath is misled, and when merging each mode adjacency information, using the means of adaptive graphics habit, for letter contained by each mode
The accounting weights of breath carry out self-adaptive estimation, reduce the inaccuracy, not scientific for manually setting and bringing.
Further, in one embodiment of the invention, existed according to the corresponding relation of original incomplete isomeric data
Each mode of the multiple mode selects anchor point respectively.
Further, in one embodiment of the invention, anchor salted hash Salted is passed through according to the unified adjacency information
Unified Hash coding is obtained, is further comprised：Anchor figure is built using the unified adjacency information；Obtain the unified Kazakhstan
Uncommon coding；Information retrieval is carried out according to the unified Hash coding.
Further, in one embodiment of the invention, described practised by adaptive graphics merges the multiple mode
The adjacency information of each mode obtain unified adjacency information, further comprise：Obtained by adaptive learning described each
The accounting weight of mode, and learn to obtain the adaptive figure, to obtain the unified adjacency information.
Further, in one embodiment of the invention, learnt according to the adaptive figure, with to described each
The accounting weights of information contained by mode carry out self-adaptive estimation, so as to obtain the adaptive figure.
To reach above-mentioned purpose, another aspect of the present invention embodiment proposes a kind of for the adaptive of imperfect isomeric data
Hash device is answered, including：Module is built, for when isomeric data sample lacks, distinguishing structure to each mode of multiple mode
Build bipartite graph and carry out Markov random walk and probability transfer calculating, and estimate the adjacency information for being related to lacking sample；
Study module, the adjacency information for practising each mode for merging the multiple mode by adaptive graphics obtain unified neighbour
Information is connect, and unified Hash is obtained by anchor salted hash Salted according to the unified adjacency information and encoded.
The adaptive Hash device for imperfect isomeric data of the embodiment of the present invention, in a unified vague generalization frame
In frame, the retrieval effectiveness of system is improved by the Hash coding techniques of linear session, wherein, it is related to lacking sample in estimation
Adjacency information when, it is only necessary to utilize existing data, it is not necessary to auxiliary information is additionally introduced, so as to reduce unnecessary letter
Breath is misled, and when merging each mode adjacency information, using the means of adaptive graphics habit, for letter contained by each mode
The accounting weights of breath carry out self-adaptive estimation, reduce the inaccuracy, not scientific for manually setting and bringing.
Further, in one embodiment of the invention, existed according to the corresponding relation of original incomplete isomeric data
Each mode of the multiple mode selects anchor point respectively.
Further, in one embodiment of the invention, the study module includes：Construction unit, for utilizing institute
State unified adjacency information structure anchor figure；Acquiring unit, for obtaining the unified Hash coding；Retrieval unit, for root
Information retrieval is carried out according to the unified Hash coding.
Further, in one embodiment of the invention, the study module is additionally operable to obtain by adaptive learning
The accounting weight of each mode, and learn to obtain the adaptive figure, to obtain the unified adjacency information.
Further, in one embodiment of the invention, learnt according to the adaptive figure, with to described each
The accounting weights of information contained by mode carry out self-adaptive estimation, so as to obtain the adaptive figure.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partly become from the following description
Obtain substantially, or recognized by the practice of the present invention.
Brief description of the drawings
Of the invention above-mentioned and/or additional aspect and advantage will become from the following description of the accompanying drawings of embodiments
Substantially and it is readily appreciated that, wherein：
Fig. 1 is the flow according to the adaptive hash method for imperfect isomeric data of one embodiment of the invention
Figure；
Fig. 2 is the stream according to the adaptive hash method for imperfect isomeric data of a specific embodiment of the invention
Cheng Tu；
Fig. 3 is to be shown according to the structure of the adaptive Hash device for imperfect isomeric data of one embodiment of the invention
It is intended to.
Embodiment
Embodiments of the invention are described below in detail, the example of the embodiment is shown in the drawings, wherein from beginning to end
Same or similar label represents same or similar element or the element with same or like function.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to for explaining the present invention, and is not considered as limiting the invention.
The adaptive Hash for imperfect isomeric data proposed according to embodiments of the present invention is described with reference to the accompanying drawings
Method and device, describe to propose according to embodiments of the present invention first with reference to the accompanying drawings is directed to the adaptive of imperfect isomeric data
Hash method.
Fig. 1 is the flow chart of the adaptive hash method for imperfect isomeric data of one embodiment of the invention.
As shown in figure 1, it should comprise the following steps for the adaptive hash method of imperfect isomeric data：
In step S101, when isomeric data sample lacks, bipartite graph is built respectively to each mode of multiple mode
Carry out Markov random walk and probability transfer calculates, and estimate the adjacency information for being related to lacking sample.
It is understood that targetedly measure is proposed for two challenges that imperfect isomeric data is brought, to carry
Retrieval performance is risen, first challenge that the method for the embodiment of the present invention faces is the Information recovering problem on lacking sample, and
For this challenge, the embodiment of the present invention carries out probability transfer calculating by building bipartite graph respectively to each mode, estimates
It is related to the adjacency information for lacking sample.
Further, in one embodiment of the invention, existed according to the corresponding relation of original incomplete isomeric data
Each mode of multiple mode selects anchor point respectively.
That is, as shown in Fig. 2 first according to the corresponding relation of original incomplete isomeric data, in each mode
Anchor point is selected respectively, secondly builds bipartite graph respectively in each mode.Markov random walk is carried out on each bipartite graph
Shifted with probability, estimate the adjacency information for being related to lacking sample.
In step s 102, the adjacency information for each mode for merging multiple mode being practised by adaptive graphics is unified
Adjacency information, and unified Hash is obtained by anchor salted hash Salted according to unified adjacency information and encoded.
It is understood that second challenge that the method for the embodiment of the present invention faces is encoded on each mode Hash
The problem of how uniformity is maintained, and for this challenge, the embodiment of the present invention is practised by adaptive graphics to be unified
Adjacency information, and then pass through anchor salted hash Salted, you can obtain unified Hash for each mode and encode.
Further, in one embodiment of the invention, each mould for merging multiple mode is practised by adaptive graphics
The adjacency information of state obtains unified adjacency information, further comprises：The accounting that each mode is obtained by adaptive learning is weighed
Weight, and learn adaptively to be schemed, to obtain unified adjacency information.
Wherein, in one embodiment of the invention, learnt according to adaptive figure, with to letter contained by each mode
The accounting weights of breath carry out self-adaptive estimation, so as to adaptively be schemed.
It is understood that the adjacency information estimated respectively for each mode, the means practised using adaptive graphics
Merged.The adaptive information proportion that should determine that each mode, and then obtain unified adjacency information.
Further, in one embodiment of the invention, obtained according to unified adjacency information by anchor salted hash Salted
Unified Hash coding, further comprises：Anchor figure is built using unified adjacency information；Obtain unified Hash coding；According to
Unified Hash coding carries out information retrieval.
It is understood that as shown in Fig. 2 carry out the structure of anchor figure using unified adjacency information, Hash is then carried out
Coding and information retrieval.
For example, user's usage scenario 1：Isomeric data sample lacks.Retrieved for large scale scale heterogeneous data, such as
Each mode sample standard deviation of fruit has missing, and preferably retrieval performance can be obtained in linear session using present invention method；
User's usage scenario 2：The affinity information completion of figure.For the isomeric data of lack part sample, the embodiment of the present invention is used
Method is estimated that the adjacency information that missing sample is related to, and then the similarity matrix of completion figure；User's usage scenario 3：
Multimodal information fusion., can be each with adaptive learning using present invention method for each mode of isomeric data
The accounting weight of mode, while can learn to obtain a unified figure.
In summary, the method for the embodiment of the present invention proposes a general Kazakhstan first against imperfect isomeric data
Uncommon coding techniques framework, effective lifting of retrieval performance, the speed of service are realized by probability transfer and adaptive information fusion
Also it is significantly improved, only linear with total sample number, the Information recovering that sample is secondly lacked for being related to each mode is asked
Topic, the present invention carry out Markov random walk and probability transfer calculating, estimation by building bipartite graph respectively to each mode
The problem of going out the adjacency information for being related to lacking sample, and how being maintained for the uniformity of each mode Hash coding,
The present invention practises the adjacency information for merging each mode by adaptive graphics, and then passes through anchor salted hash Salted, you can for each mode
Obtain unified Hash coding.
The adaptive hash method for imperfect isomeric data according to embodiments of the present invention is unified general at one
Change in framework, the retrieval effectiveness of system is improved by the Hash coding techniques of linear session, wherein, it is related to lacking in estimation
During the adjacency information of sample, it is only necessary to utilize existing data, it is not necessary to auxiliary information is additionally introduced, it is unnecessary so as to reduce
Information misguidance, and when merging each mode adjacency information, the means practised using adaptive graphics, for contained by each mode
The accounting weights for having information carry out self-adaptive estimation, reduce the inaccuracy, not scientific for manually setting and bringing, and it is for solving
Certainly in isomeric data in the case of each equal lack part sample of mode, effectively solve to be related to the Information recovering for lacking sample
The problem of how problem and the uniformity of each mode Hash coding keep, effectively lift the information retrieval performance of salted hash Salted.
The adaptive Hash for imperfect isomeric data proposed according to embodiments of the present invention referring next to accompanying drawing description
Device.
Fig. 3 is the structural representation of the adaptive Hash device for imperfect isomeric data of one embodiment of the invention
Figure.
As shown in figure 3, it should include for the adaptive Hash device 10 of imperfect isomeric data：Build module 100 and learn
Practise module 200.
Wherein, build module 100 to be used for when isomeric data sample lacks, each mode of multiple mode is built respectively
Bipartite graph carries out Markov random walk and probability transfer calculates, and estimates the adjacency information for being related to lacking sample.Learn
Practise module 200 and obtain unified adjacent letter for practising the adjacency information for each mode for merging multiple mode by adaptive graphics
Breath, and unified Hash is obtained by anchor salted hash Salted according to unified adjacency information and encoded.The device 10 of the embodiment of the present invention
The information retrieval performance of salted hash Salted can be effectively lifted, obtaining unified Hash by study encodes, so as to lifting system
Retrieval effectiveness, reduces unnecessary information misguidance, reduces the inaccuracy, not scientific for manually setting and bringing.
Further, in one embodiment of the invention, existed according to the corresponding relation of original incomplete isomeric data
Each mode of multiple mode selects anchor point respectively.
Further, in one embodiment of the invention, study module 200 includes：Construction unit, acquiring unit and inspection
Cable elements.
Wherein, construction unit is used to utilize unified adjacency information structure anchor figure.Acquiring unit is used to obtain unified Kazakhstan
Uncommon coding.Retrieval unit is used to carry out information retrieval according to unified Hash coding.
Further, in one embodiment of the invention, study module 200 is additionally operable to obtain often by adaptive learning
The accounting weight of individual mode, and learn adaptively to be schemed, to obtain unified adjacency information.
Further, in one embodiment of the invention, learnt according to adaptive figure, with to contained by each mode
The accounting weights for having information carry out self-adaptive estimation, so as to adaptively be schemed.
It should be noted that the explanation of the foregoing adaptive hash method embodiment to for imperfect isomeric data
The adaptive Hash device for imperfect isomeric data of the embodiment is also applied for, here is omitted.
The adaptive Hash device for imperfect isomeric data according to embodiments of the present invention is unified general at one
Change in framework, the retrieval effectiveness of system is improved by the Hash coding techniques of linear session, wherein, it is related to lacking in estimation
During the adjacency information of sample, it is only necessary to utilize existing data, it is not necessary to auxiliary information is additionally introduced, it is unnecessary so as to reduce
Information misguidance, and when merging each mode adjacency information, the means practised using adaptive graphics, for contained by each mode
The accounting weights for having information carry out self-adaptive estimation, reduce the inaccuracy, not scientific for manually setting and bringing, and it is for solving
Certainly in isomeric data in the case of each equal lack part sample of mode, effectively solve to be related to the Information recovering for lacking sample
The problem of how problem and the uniformity of each mode Hash coding keep, effectively lift the information retrieval performance of salted hash Salted.
In the description of the invention, it is to be understood that term " " center ", " longitudinal direction ", " transverse direction ", " length ", " width ",
" thickness ", " on ", " under ", "front", "rear", "left", "right", " vertical ", " level ", " top ", " bottom " " interior ", " outer ", " up time
The orientation or position relationship of the instruction such as pin ", " counterclockwise ", " axial direction ", " radial direction ", " circumference " be based on orientation shown in the drawings or
Position relationship, it is for only for ease of and describes the present invention and simplify description, rather than indicates or imply that signified device or element must
There must be specific orientation, with specific azimuth configuration and operation, therefore be not considered as limiting the invention.
In addition, term " first ", " second " are only used for describing purpose, and it is not intended that instruction or hint relative importance
Or the implicit quantity for indicating indicated technical characteristic.Thus, define " first ", the feature of " second " can be expressed or
Implicitly include at least one this feature.In the description of the invention, " multiple " are meant that at least two, such as two, three
It is individual etc., unless otherwise specifically defined.
In the present invention, unless otherwise clearly defined and limited, term " installation ", " connected ", " connection ", " fixation " etc.
Term should be interpreted broadly, for example, it may be fixedly connected or be detachably connected, or integrally；Can be that machinery connects
Connect or electrically connect；Can be joined directly together, can also be indirectly connected by intermediary, can be in two elements
The connection in portion or the interaction relationship of two elements, limited unless otherwise clear and definite.For one of ordinary skill in the art
For, the concrete meaning of above-mentioned term in the present invention can be understood as the case may be.
In the present invention, unless otherwise clearly defined and limited, fisrt feature can be with "above" or "below" second feature
It is that the first and second features directly contact, or the first and second features pass through intermediary mediate contact.Moreover, fisrt feature exists
Second feature " on ", " top " and " above " but fisrt feature are directly over second feature or oblique upper, or be merely representative of
Fisrt feature level height is higher than second feature.Fisrt feature second feature " under ", " lower section " and " below " can be
One feature is immediately below second feature or obliquely downward, or is merely representative of fisrt feature level height and is less than second feature.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or the spy for combining the embodiment or example description
Point is contained at least one embodiment or example of the present invention.In this manual, to the schematic representation of above-mentioned term not
Identical embodiment or example must be directed to.Moreover, specific features, structure, material or the feature of description can be with office
Combined in an appropriate manner in one or more embodiments or example.In addition, in the case of not conflicting, the skill of this area
Art personnel can be tied the different embodiments or example and the feature of different embodiments or example described in this specification
Close and combine.
Although embodiments of the invention have been shown and described above, it is to be understood that above-described embodiment is example
Property, it is impossible to limitation of the present invention is interpreted as, one of ordinary skill in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, changed, replacing and modification.
Claims (10)
1. a kind of adaptive hash method for imperfect isomeric data, it is characterised in that comprise the following steps：
When isomeric data sample lacks, build bipartite graph respectively to each mode of multiple mode and carry out Markov and swim at random
Walk to shift with probability and calculate, and estimate the adjacency information for being related to lacking sample；
The adjacency information that each mode for merging the multiple mode is practised by adaptive graphics obtains unified adjacency information, and
Unified Hash is obtained according to the unified adjacency information by anchor salted hash Salted to encode.
2. the adaptive hash method according to claim 1 for imperfect isomeric data, it is characterised in that according to original
The begin corresponding relation of incomplete isomeric data selects anchor point respectively in each mode of the multiple mode.
3. the adaptive hash method according to claim 1 for imperfect isomeric data, it is characterised in that according to institute
State unified adjacency information and unified Hash coding is obtained by anchor salted hash Salted, further comprise：
Anchor figure is built using the unified adjacency information；
Obtain the unified Hash coding；And
Information retrieval is carried out according to the unified Hash coding.
4. the adaptive hash method according to claim 1 for imperfect isomeric data, it is characterised in that described logical
The adjacency information for crossing each mode that adaptive graphics practises the multiple mode of fusion obtains unified adjacency information, further bag
Include：
The accounting weight of each mode is obtained by adaptive learning, and learns to obtain the adaptive figure, to obtain
State unified adjacency information.
5. the adaptive hash method according to claim 4 for imperfect isomeric data, it is characterised in that wherein,
Learnt according to the adaptive figure, self-adaptive estimation carried out with the accounting weights to information contained by each mode,
So as to obtain the adaptive figure.
A kind of 6. adaptive Hash device for imperfect isomeric data, it is characterised in that including：
Module is built, for when isomeric data sample lacks, bipartite graph progress to be built respectively to each mode of multiple mode
Markov random walk and probability transfer calculate, and estimate the adjacency information for being related to lacking sample；
Study module, the adjacency information for practising each mode for merging the multiple mode by adaptive graphics are unified
Adjacency information, and unified Hash is obtained by anchor salted hash Salted according to the unified adjacency information and encoded.
7. the adaptive Hash device according to claim 6 for imperfect isomeric data, it is characterised in that according to original
The begin corresponding relation of incomplete isomeric data selects anchor point respectively in each mode of the multiple mode.
8. the adaptive Hash device according to claim 6 for imperfect isomeric data, it is characterised in that
Practising module includes：
Construction unit, for building anchor figure using the unified adjacency information；
Acquiring unit, for obtaining the unified Hash coding；And
Retrieval unit, for carrying out information retrieval according to the unified Hash coding.
9. the adaptive Hash device according to claim 6 for imperfect isomeric data, it is characterised in that
Practise module to be additionally operable to obtain the accounting weight of each mode by adaptive learning, and learn to obtain the adaptive figure,
To obtain the unified adjacency information.
10. the adaptive Hash device according to claim 9 for imperfect isomeric data, it is characterised in that wherein,
Learnt according to the adaptive figure, self-adaptive estimation carried out with the accounting weights to information contained by each mode,
So as to obtain the adaptive figure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710544048.2A CN107391619A (en) | 2017-07-05 | 2017-07-05 | For the adaptive hash method and device of imperfect isomeric data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710544048.2A CN107391619A (en) | 2017-07-05 | 2017-07-05 | For the adaptive hash method and device of imperfect isomeric data |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107391619A true CN107391619A (en) | 2017-11-24 |
Family
ID=60335267
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710544048.2A Pending CN107391619A (en) | 2017-07-05 | 2017-07-05 | For the adaptive hash method and device of imperfect isomeric data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107391619A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111090765A (en) * | 2019-11-25 | 2020-05-01 | 山东师范大学 | Social image retrieval method and system based on missing multi-modal hash |
| CN111104599A (en) * | 2019-12-23 | 2020-05-05 | 北京百度网讯科技有限公司 | Method and apparatus for outputting information |
| CN111160426A (en) * | 2019-12-17 | 2020-05-15 | 齐鲁工业大学 | Feature fusion method and system based on tensor fusion and LSTM network |
| CN117155583A (en) * | 2023-10-24 | 2023-12-01 | 清华大学 | Multi-mode identity authentication method and system for incomplete information deep fusion |
-
2017
- 2017-07-05 CN CN201710544048.2A patent/CN107391619A/en active Pending
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111090765A (en) * | 2019-11-25 | 2020-05-01 | 山东师范大学 | Social image retrieval method and system based on missing multi-modal hash |
| CN111090765B (en) * | 2019-11-25 | 2020-09-29 | 山东师范大学 | Social image retrieval method and system based on missing multi-modal hash |
| CN111160426A (en) * | 2019-12-17 | 2020-05-15 | 齐鲁工业大学 | Feature fusion method and system based on tensor fusion and LSTM network |
| CN111160426B (en) * | 2019-12-17 | 2023-04-28 | 齐鲁工业大学 | Feature fusion method and system based on tensor fusion and LSTM network |
| CN111104599A (en) * | 2019-12-23 | 2020-05-05 | 北京百度网讯科技有限公司 | Method and apparatus for outputting information |
| CN111104599B (en) * | 2019-12-23 | 2023-08-18 | 北京百度网讯科技有限公司 | Method and device for outputting information |
| CN117155583A (en) * | 2023-10-24 | 2023-12-01 | 清华大学 | Multi-mode identity authentication method and system for incomplete information deep fusion |
| CN117155583B (en) * | 2023-10-24 | 2024-01-23 | 清华大学 | Multi-mode identity authentication method and system for incomplete information deep fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107391619A (en) | For the adaptive hash method and device of imperfect isomeric data | |
| CN109493346B (en) | Stomach cancer pathological section image segmentation method and device based on multiple losses | |
| CN110415516A (en) | Urban traffic flow prediction technique and medium based on figure convolutional neural networks | |
| CN107066445B (en) | The deep learning method of one attribute emotion word vector | |
| CN108509408A (en) | A kind of sentence similarity judgment method | |
| EP3929824A3 (en) | Robust multimodal sensor fusion for autonomous driving vehicles | |
| CN108665506A (en) | Image processing method, device, computer storage media and server | |
| CN110222770A (en) | A kind of vision answering method based on syntagmatic attention network | |
| CN106295245A (en) | The method of storehouse noise reduction own coding gene information feature extraction based on Caffe | |
| CN106409024A (en) | Children' adaptive ability improvement system and method | |
| CN109598298A (en) | Image object recognition methods and system | |
| CN110223553B (en) | Method and system for predicting answer information | |
| CN113255701B (en) | Small sample learning method and system based on absolute-relative learning framework | |
| CN107452374A (en) | Based on unidirectional from the various visual angles language identification method for marking auxiliary information | |
| CN113159052B (en) | Method for identifying failure mode of flexural reinforced concrete simply supported beam based on deep learning | |
| CN110175627A (en) | A kind of complicated classification method of knowledge based recombination | |
| CN110852808A (en) | Asynchronous adaptive value evaluation method of electronic product based on deep neural network | |
| CN105894407A (en) | Method for evaluating learning ability development trend of students | |
| CN109933731A (en) | A kind of friend recommendation method, apparatus, equipment and storage medium | |
| CN116432750B (en) | Box embedding-based small sample knowledge graph completion method | |
| CN106251307A (en) | The image approximate method converted based on Extended Smoothlets (ExSmoothlets) | |
| CN116070105A (en) | Combined beam damage identification method and system based on wavelet transformation and residual error network | |
| CN106203266B (en) | The extracting method and device of image extreme point | |
| CN105512626B (en) | A kind of method of numeral shown by identification charactron | |
| CN112232350B (en) | Paddy field robot mechanical leg length adjusting method and system based on reinforcement learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171124 |