CN117391646B - Collaborative innovation management system - Google Patents
Collaborative innovation management system Download PDFInfo
- Publication number
- CN117391646B CN117391646B CN202311685364.3A CN202311685364A CN117391646B CN 117391646 B CN117391646 B CN 117391646B CN 202311685364 A CN202311685364 A CN 202311685364A CN 117391646 B CN117391646 B CN 117391646B
- Authority
- CN
- China
- Prior art keywords
- feature
- training
- enterprise
- government
- matrix
- 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.)
- Active
Links
- 239000011159 matrix material Substances 0.000 claims abstract description 211
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000009826 distribution Methods 0.000 claims abstract description 17
- 230000008569 process Effects 0.000 claims abstract description 15
- 238000012937 correction Methods 0.000 claims abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims abstract description 11
- 238000012549 training Methods 0.000 claims description 135
- 239000013598 vector Substances 0.000 claims description 110
- 238000013527 convolutional neural network Methods 0.000 claims description 74
- 230000008901 benefit Effects 0.000 claims description 26
- 238000012545 processing Methods 0.000 claims description 16
- 238000010586 diagram Methods 0.000 claims description 14
- 230000004913 activation Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 7
- 238000012546 transfer Methods 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 238000011156 evaluation Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 238000011161 development Methods 0.000 claims description 4
- 238000012827 research and development Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 230000001902 propagating effect Effects 0.000 claims description 2
- 230000009467 reduction Effects 0.000 claims description 2
- 238000011160 research Methods 0.000 abstract description 15
- 238000003062 neural network model Methods 0.000 abstract description 7
- 238000007726 management method Methods 0.000 description 35
- 238000004458 analytical method Methods 0.000 description 6
- 239000000284 extract Substances 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000000875 corresponding effect Effects 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 102100033620 Calponin-1 Human genes 0.000 description 1
- 102100033591 Calponin-2 Human genes 0.000 description 1
- 102100033592 Calponin-3 Human genes 0.000 description 1
- 101000945318 Homo sapiens Calponin-1 Proteins 0.000 description 1
- 101000945403 Homo sapiens Calponin-2 Proteins 0.000 description 1
- 101000945410 Homo sapiens Calponin-3 Proteins 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000012550 audit Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
Classifications
-
- 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
- G06Q10/00—Administration; Management
- G06Q10/10—Office automation; Time management
- G06Q10/101—Collaborative creation, e.g. joint development of products or services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/2433—Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
Abstract
The utility model discloses a collaborative innovation management system, which is characterized in that global implicit associated feature distribution information of all parameters in government parameters, enterprise parameters and college parameters in the collaborative innovation system is respectively extracted through a deep neural network model, the feature information of the three are fused by using a convolution neural network with a three-dimensional convolution kernel, so that dynamic implicit associated features of the three with deeper layers are extracted, and the feature matrix of the three is further subjected to Cauchy weight probability correction in the process, so that the robustness of information loss is enhanced, and the feature expression capability of a three-party associated feature map fused with the feature information is improved. Thus, the efficiency of innovation in obstetric and research can be improved.
Description
Technical Field
The invention relates to the field of intelligent collaborative management, and more particularly relates to a collaborative innovation management system.
Background
The obstetric and scientific research collaborative innovation is taken as a brand new organization mode for improving the autonomous innovation capability of countries and regions, becomes a new trend of the current international technological innovation activities and a new focus of innovation theory research, can realize the mutual coupling of all links from knowledge production to knowledge commercialization, and is a primary choice for solving the problem of loose connection of education, science and technology and economic society development. In the process of government audit, the research on the collaborative innovation of obstetric and academic research at home and abroad is concentrated on theoretical analysis and mechanism research, a three-party game system based on the condition that government, enterprise and university parties are all collaborative innovation game main bodies is not constructed, and the change of the collaborative innovation relation can not be simulated under the condition of different factors. Therefore, the construction of the three-party evolution game analysis system for political production and research has important significance in both theory and reality.
Most researches take enterprises and universities as game main bodies of collaborative innovation, only government actions are introduced into a game model as exogenous variables, and the government is not taken as a behavior main body to carry out game analysis with the enterprises and the universities; the collaborative innovation three-party game model of the college draw in collaborative innovation of the political obstetrics and research is not researched, the detailed analysis of the mode of participating in collaborative innovation by the government is not performed, only fund support is often considered, and the policy support which is more important in reality is ignored, so that the benefits brought to enterprises and universities are omitted. And meanwhile, the game between the cooperative innovation main bodies does not consider the influence of participation willingness on the selection of the strategy. Thus, the efficiency of innovation of the obstetric and research cooperation is greatly reduced, and the development of the obstetric and research cooperation is not facilitated. Thus, there is a need for a collaborative innovation management system.
Disclosure of Invention
The present application has been made in order to solve the above technical problems. The embodiment of the application provides a collaborative innovation management system, which is characterized in that global implicit associated feature distribution information of government parameters, enterprise parameters and college parameters in the collaborative innovation system is respectively extracted through a deep neural network model, the feature information of the three are fused by using a convolution neural network with a three-dimensional convolution kernel, so that dynamic implicit associated features of the three are extracted, and the feature matrix of the three is further subjected to Cauchy weight probability correction in the process, so that the robustness to information loss is enhanced, and the feature expression capability of a three-party associated feature map of the fused feature information is improved. Thus, the efficiency of innovation in obstetric and research can be improved.
According to one aspect of the present application, there is provided a collaborative innovation management system comprising:
the innovation parameter acquisition unit is used for acquiring a plurality of first parameters of a government, a plurality of second parameters of an enterprise and a plurality of third parameters of a university in the collaborative innovation system;
the parameter semantic coding unit is used for respectively passing the first parameters of the government, the second parameters of the enterprise and the third parameters of the university through a context encoder comprising an embedded layer to obtain a plurality of government parameter feature vectors, a plurality of enterprise parameter feature vectors and a plurality of university parameter feature vectors;
the government parameter association coding unit is used for two-dimensionally arranging the government parameter feature vectors into a first feature matrix and then obtaining the government feature matrix through a first convolutional neural network;
the enterprise parameter association coding unit is used for two-dimensionally arranging the plurality of enterprise parameter feature vectors into a second feature matrix and then obtaining an enterprise feature matrix through a second convolutional neural network;
the college parameter association coding unit is used for two-dimensionally arranging the plurality of college parameter feature vectors into a third feature matrix and then obtaining the college feature matrix through a third convolutional neural network;
And the characteristic distribution correction unit is used for carrying out cauchy re-probability on the government characteristic matrix, the enterprise characteristic matrix and the college characteristic matrix to obtain a corrected government characteristic matrix, a corrected enterprise characteristic matrix and a corrected college characteristic matrix, wherein the Ke Xichong probability is carried out based on the proportion between the characteristic value of each position in the characteristic matrix and the sum value of the characteristic values of all positions in the characteristic matrix.
The three-party association coding unit is used for constructing the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix into three-dimensional input tensors and then obtaining a three-party association feature map through a fourth convolution neural network with a three-dimensional convolution kernel; and the innovation mode evaluation result generation unit is used for enabling the three-party associated feature graphs to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the collaborative innovation modes of governments, enterprises and universities in the collaborative innovation system are abnormal or not.
Compared with the prior art, the collaborative innovation management system provided by the application extracts global implicit associated feature distribution information of government parameters, enterprise parameters and college parameters in the collaborative innovation system respectively through the deep neural network model, and uses the convolution neural network with the three-dimensional convolution kernel to fuse the feature information of the three to extract dynamic implicit associated features of the three in a deeper level, and further carries out Cauchy weight probability correction on the feature matrixes of the three in the process so as to enhance the robustness to information loss and improve the feature expression capability of the three-party associated feature map fused with the feature information. Thus, the efficiency of innovation in obstetric and research can be improved.
Drawings
The foregoing and other objects, features and advantages of the present application will become more apparent from the following more particular description of embodiments of the present application, as illustrated in the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the application and are incorporated in and constitute a part of this specification, illustrate the application and not constitute a limitation to the application. In the drawings, like reference numerals generally refer to like parts or steps.
Fig. 1 is a block diagram of a collaborative innovation management system according to an embodiment of the present application.
Fig. 2 is a block diagram of a parameter semantic coding unit in a collaborative innovation management system according to an embodiment of the application.
Fig. 3 is a flowchart of a management method of the collaborative innovation management system according to an embodiment of the present application.
Fig. 4 is a schematic architecture diagram of a management method of a collaborative innovation management system according to an embodiment of the application.
Description of the embodiments
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some of the embodiments of the present application and not all of the embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
Fig. 1 illustrates a block diagram of a collaborative innovation management system according to an embodiment of the present application. As shown in fig. 1, a collaborative innovation management system 200 according to an embodiment of the present application includes: a innovation parameter obtaining unit 210 for obtaining a plurality of first parameters of a government, a plurality of second parameters of an enterprise, and a plurality of third parameters of a university in a collaborative innovation system; a parameter semantic coding unit 220, configured to obtain a plurality of government parameter feature vectors, a plurality of enterprise parameter feature vectors and a plurality of university parameter feature vectors by respectively passing the plurality of government first parameters, the plurality of second parameters of the enterprise and the plurality of third parameters of the university through a context encoder including an embedded layer; the government parameter association coding unit 230 is configured to two-dimensionally arrange the plurality of government parameter feature vectors into a first feature matrix, and then obtain a government feature matrix through a first convolutional neural network; the enterprise parameter association coding unit 240 is configured to two-dimensionally arrange the plurality of enterprise parameter feature vectors into a second feature matrix, and then obtain an enterprise feature matrix through a second convolutional neural network; the college parameter association coding unit 250 is configured to two-dimensionally arrange the plurality of college parameter feature vectors into a third feature matrix, and then obtain a college feature matrix through a third convolutional neural network; a feature distribution correction unit 260, configured to perform cauchy weight probability on the government feature matrix, the enterprise feature matrix, and the college feature matrix to obtain a corrected government feature matrix, a corrected enterprise feature matrix, and a corrected college feature matrix, where the Ke Xichong probability is performed based on a ratio between a feature value of each position in the feature matrix and a sum value of feature values of all positions in the feature matrix; the three-party association encoding unit 270 is configured to construct the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix into three-dimensional input tensors, and then obtain a three-party association feature map through a fourth convolutional neural network with a three-dimensional convolutional kernel; and an innovation pattern evaluation result generating unit 280, configured to pass the three-party associated feature graphs through a classifier to obtain a classification result, where the classification result is used to indicate whether there is an abnormality in the collaborative innovation patterns of the government, the enterprise and the university in the collaborative innovation system.
Specifically, in the embodiment of the present application, the innovating party parameter obtaining unit 210 and the parameter semantic encoding unit 220 are configured to obtain a plurality of first parameters of a government, a plurality of second parameters of an enterprise, and a plurality of third parameters of a university in a collaborative innovation system, and pass the plurality of first parameters of the government, the plurality of second parameters of the enterprise, and the plurality of third parameters of the university through a context encoder including an embedded layer, respectively, to obtain a plurality of government parameter feature vectors, a plurality of enterprise parameter feature vectors, and a plurality of university parameter feature vectors. As described above, conventional collaborative innovation management analysis is a statistical-based analysis, but since a three-party game system composed of government, enterprise and university is a complex system, it has nonlinear and unbalanced essential characteristics, and cannot be accurately evaluated by a conventional statistical model. Therefore, in the technical scheme of the application, the neural network model of deep learning is expected to be used for accurately estimating the neural network model, so that the efficiency of innovation of obstetrical research and development cooperation is improved.
That is, specifically, in the technical solution of the present application, first, a plurality of first parameters of a government, a plurality of second parameters of an enterprise, and a plurality of third parameters of a university in a collaborative innovation system are acquired. In particular, herein, the first plurality of government parameters includes a total cost of government participation in the collaborative innovation, a benefit obtained by government selection "participation" policy, a proportion of benefit obtained by government selection "non-participation" policy to benefit obtained by government selection "participation" policy, and a willingness of government selection to participate in the collaborative innovation; the second parameters of the enterprise include a reduction amount of total cost input by the enterprise and the university in the collaborative innovation process by a preferential policy provided by the government, initial benefits before the enterprise performs the collaborative innovation, benefits obtained by independent research and development of the enterprise, punishments paid to the enterprise when the enterprise breaks down, willingness of the enterprise to select the collaborative innovation, additional benefits brought to the enterprise when the enterprise and the university select the collaborative innovation, total cost of the enterprise and the university participating in the collaborative innovation and cost sharing proportionality coefficients of the enterprise; and, the plurality of third parameters of the university include a governmental provided benefit policy such that the enterprise and the university are reduced in total cost invested in the collaborative innovation process, an initial benefit before the collaborative innovation is performed by the university, a benefit obtained by the independent development of the university, a penalty paid to the enterprise when the university breaks, a willingness of the university to choose to perform the collaborative innovation, an additional benefit brought to the university when the enterprise and the university both choose to perform the collaborative innovation, a total cost of the enterprise and the university to participate in the collaborative innovation, a fund support given by the government to the university with the active parameter collaborative innovation, and a cost allocation scaling factor of the university.
The first plurality of government parameters are then processed through a global information encoding process in a context encoder comprising an embedded layer to obtain a plurality of government parameter feature vectors having global government parameter implicitly associated feature information. And, for the second parameters of the enterprise and the third parameters of the university, respectively, the second parameters and the third parameters of the university are also encoded in a context encoder comprising an embedded layer, so as to respectively extract a plurality of enterprise parameter feature vectors with global enterprise parameter implicit associated feature information and extract a plurality of university parameter feature vectors with global university parameter implicit associated feature information.
More specifically, in an embodiment of the present application, the parameter semantic coding unit includes: an input vector construction subunit for constructing a plurality of first parameters of the government, a plurality of second parameters of the enterprise, and a plurality of third parameters of the university as input vectors, respectively, to obtain a sequence of input vectors; an embedding subunit for converting each input vector in the sequence of input vectors into an embedded vector using an embedding layer of the context encoder including an embedding layer to obtain a sequence of embedded vectors; and a context semantic coding subunit configured to perform global-based context semantic coding on the sequence of embedded vectors using a translator of the context encoder including an embedded layer to obtain the plurality of government parameter feature vectors, the plurality of enterprise parameter feature vectors, and the plurality of university parameter feature vectors.
Fig. 2 illustrates a block diagram of a parameter semantic coding unit in a collaborative innovation management system according to an embodiment of the application. As shown in fig. 2, the parameter semantic coding unit 220 includes: an input vector construction subunit 221 configured to construct a plurality of first parameters of the government, a plurality of second parameters of the enterprise, and a plurality of third parameters of the university as input vectors, respectively, to obtain a sequence of input vectors; an embedding subunit 222, configured to convert each input vector in the sequence of input vectors into an embedded vector using an embedding layer of the context encoder that includes an embedding layer to obtain a sequence of embedded vectors; and a context semantic coding subunit 223 for globally based context semantic coding the sequence of embedded vectors using the translator of the context encoder comprising an embedded layer to obtain the plurality of government parameter feature vectors, the plurality of enterprise parameter feature vectors, and the plurality of college parameter feature vectors.
Specifically, in the embodiment of the present application, the government parameter association encoding unit 230 is configured to two-dimensionally arrange the plurality of government parameter feature vectors into a first feature matrix, and then obtain the government feature matrix through a first convolutional neural network. That is, in the technical solution of the present application, further, the government parameter feature vectors are two-dimensionally arranged into a first feature matrix, and then the first feature matrix is obtained through a first convolutional neural network. In this way, after integrating the correlation features among the plurality of government parameter feature information, the feature information of the plurality of government parameters may be subjected to a deeper feature mining using a convolutional neural network model having excellent performance in terms of high-dimensional implicit correlation feature extraction to obtain the government feature matrix.
More specifically, in the embodiment of the present application, the government parameter-related encoding unit is further configured to use each layer of the first convolutional neural network to perform, during forward transfer of the layer, input data separately: carrying out convolution processing on input data to obtain a convolution characteristic diagram; carrying out local channel dimension-based mean pooling on the convolution feature map to obtain a pooled feature map; performing nonlinear activation on the pooled feature map to obtain an activated feature map; wherein the output of the last layer of the first convolutional neural network is the government feature matrix, and the input of the first layer of the first convolutional neural network is the first feature matrix.
Specifically, in this embodiment of the present application, the enterprise parameter association encoding unit 240 and the college parameter association encoding unit 250 are configured to two-dimensionally arrange the plurality of enterprise parameter feature vectors into a second feature matrix, and then pass through a second convolutional neural network to obtain an enterprise feature matrix, and two-dimensionally arrange the plurality of college parameter feature vectors into a third feature matrix, and then pass through a third convolutional neural network to obtain a college feature matrix. In other words, in the technical solution of the present application, further, by two-dimensionally arranging the plurality of enterprise parameter feature vectors into the second feature matrix in the same manner, then processing the second feature matrix in the second convolutional neural network, and two-dimensionally arranging the plurality of university parameter feature vectors into the third feature matrix, then processing the third feature matrix in the third convolutional neural network, an enterprise feature matrix corresponding to an enterprise and having global enterprise parameter data associated feature information and a university feature matrix corresponding to a university and having global university parameter data associated feature information can be obtained. Accordingly, in one specific example, each layer of the third convolutional neural network is used to perform convolution processing, mean pooling processing of local channel dimensions and nonlinear activation processing on input data during forward transfer of the layer, respectively, so as to output the college feature matrix by the last layer of the third convolutional neural network.
More specifically, in the embodiment of the present application, the enterprise parameter association encoding unit is further configured to use each layer of the second convolutional neural network to perform, during forward transfer of the layer, input data separately: carrying out convolution processing on input data to obtain a convolution characteristic diagram; carrying out local channel dimension-based mean pooling on the convolution feature map to obtain a pooled feature map; performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the second convolutional neural network is the enterprise feature matrix, and the input of the first layer of the second convolutional neural network is the second feature matrix.
Specifically, in the embodiment of the present application, the feature distribution correction unit 260 is configured to perform cauchy weight probabilization on the government feature matrix, the enterprise feature matrix, and the college feature matrix to obtain a corrected government feature matrix, a corrected enterprise feature matrix, and a corrected college feature matrix, where the Ke Xichong probabilization is performed based on a ratio between a feature value of each position in the feature matrix and an added value of feature values of all positions in the feature matrix. It should be understood that in the technical solution of the present application, the government parameter feature, the enterprise parameter feature and the university parameter feature may be extracted by further constructing the government feature matrix, the enterprise feature matrix and the university feature matrix as three-dimensional input tensors and then processing the three-dimensional input tensors in a fourth convolutional neural network having a three-dimensional convolutional kernel. However, when the three-dimensional input tensor passes through the fourth convolutional neural network with the three-dimensional convolutional kernel to obtain the three-party correlation feature map, while the three-dimensional convolutional kernel performs correlation feature extraction on the government feature matrix, the enterprise feature matrix and the college feature matrix, there is information loss of the government feature matrix, the enterprise feature matrix and the college feature matrix themselves in an iterative process, so that compensation for this is desired. Specifically, in the technical scheme of the application, the cauchy weight probability is further performed on the government feature matrix, the enterprise feature matrix and the college feature matrix for correction.
More specifically, in an embodiment of the present application, the feature distribution correction unit is further configured to: and performing cauchy re-probability on the government feature matrix, the enterprise feature matrix and the college feature matrix respectively according to the following formulas to obtain the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix.
Wherein, the formula is:
wherein the method comprises the steps ofFeature values for respective positions of each of the government feature matrix, the enterprise feature matrix, and the college feature matrix, and +.>Representing the summation of eigenvalues for all locations of the eigenvalue matrix,exponential operation representing a matrix, said matrixThe exponent operation of (a) represents calculating a natural exponent function value raised to a power by the eigenvalues of each position in the matrix, and the eigenvalues divided by the parameter represents dividing by the parameters by the eigenvalues of each position in the eigenvalues, respectively.
Specifically, in the embodiment of the present application, the three-party association encoding unit 270 is configured to construct the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix into three-dimensional input tensors, and then obtain the three-party association feature map through a fourth convolutional neural network with a three-dimensional convolutional kernel. That is, in the technical solution of the present application, the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix are then constructed into a three-dimensional input tensor, and then a three-dimensional correlation feature map is obtained through a fourth convolutional neural network with a three-dimensional convolutional kernel. It should be appreciated that, in this way, the cauchy re-probability process performs probabilistic information interpretation on the eigenvalues of the government eigenvalue matrix, the enterprise eigenvalue matrix and the college eigenvalue matrix, so that when training one of the fourth convolutional neural network and the first to third convolutional neural networks through gradient back propagation, the parameters of one of the fourth convolutional neural network and the first to third convolutional neural networks in cascade can enable the information loss caused by feature extraction of the three-dimensional convolutional kernel to be self-adapted along with iteration, thereby enhancing the robustness on the information loss and improving the feature expression capability of the three-party correlation eigenvectors.
More specifically, in the embodiment of the present application, the three-party association encoding unit is further configured to: the fourth convolutional neural network with the three-dimensional convolutional kernel performs the respective processing on the input data in the forward transfer of the layer: performing three-dimensional convolution processing on the input data based on the three-dimensional convolution check to obtain a convolution characteristic diagram; carrying out mean pooling treatment based on a local feature matrix on the convolution feature map to obtain a pooled feature map; performing nonlinear activation on the pooled feature map to obtain an activated feature map; the output of the last layer of the fourth convolutional neural network is the three-party association feature map, and the input of the first layer of the fourth convolutional neural network is the three-dimensional input tensor.
Specifically, in the embodiment of the present application, the innovation mode evaluation result generating unit 280 is configured to pass the three-party associated feature map through a classifier to obtain a classification result, where the classification result is used to indicate whether the collaborative innovation mode of the government, the enterprise and the university in the collaborative innovation system is abnormal. That is, further, the three-party association feature diagram is passed through a classifier to obtain a classification result for indicating whether there is abnormality in the collaborative innovation mode of the government, the enterprise and the university in the collaborative innovation system. In one specific example, the classifier processes the three-party associated feature map to generate a classification result with the following formula: WhereinRepresenting that the three-party association feature map is projected as a three-party association feature vector,/a>To->Weight matrix for all connection layers of each layer, < ->To->Representing the bias matrix for each fully connected layer.
Specifically, in the embodiment of the present application, the collaborative innovation management system 200 further includes: training means for training the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the fourth convolutional neural network with three-dimensional convolutional kernels, and the classifier; wherein, training module includes: the training innovation parameter acquisition unit is used for acquiring training data, wherein the training data comprises a plurality of training first parameters of a government, a plurality of training second parameters of an enterprise and a plurality of training third parameters of a college in a collaborative innovation system; the training parameter semantic coding unit is used for respectively enabling a plurality of training first parameters of the government, a plurality of training second parameters of the enterprise and a plurality of training third parameters of the college to pass through a context encoder comprising an embedded layer to obtain a plurality of training government parameter feature vectors, a plurality of training enterprise parameter feature vectors and a plurality of training college parameter feature vectors; the training government parameter association coding unit is used for two-dimensionally arranging the training government parameter feature vectors into a training first feature matrix and then obtaining a training government feature matrix through the first convolutional neural network; the training enterprise parameter association coding unit is used for performing two-dimensional arrangement on the plurality of training enterprise parameter feature vectors to obtain a training enterprise feature matrix through the second convolutional neural network after training the second feature matrix; the training college parameter association coding unit is used for performing two-dimensional arrangement on the plurality of training college parameter feature vectors to form a training third feature matrix, and then obtaining the training college feature matrix through the third convolutional neural network; the training feature distribution correction unit is used for carrying out cauchy weight probability on the training government feature matrix, the training enterprise feature matrix and the training college feature matrix to obtain a training corrected government feature matrix, a training corrected enterprise feature matrix and a training corrected college feature matrix; the training three-party association coding unit is used for constructing the training corrected government feature matrix, the training corrected enterprise feature matrix and the training corrected college feature matrix into training three-dimensional input tensor and then obtaining a training three-party association feature map through the fourth convolution neural network with the three-dimensional convolution kernel; the classification loss calculation unit is used for enabling the training three-party association feature images to pass through the classifier to obtain classification loss function values; and the back propagation training unit is used for training the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the fourth convolutional neural network with a three-dimensional convolutional kernel and the classifier by propagating the classification loss function value in the gradient descending direction, wherein training optimization is carried out on training three-party association feature vectors obtained by projection of the training three-party association feature map during each iteration.
Particularly, in the technical scheme of the application, when the government feature matrix after training and correcting, the enterprise feature matrix after training and correcting and the college feature matrix after training and correcting are constructed as the three-dimensional input tensor and pass through the fourth convolution neural network with the three-dimensional convolution kernel, the fourth convolution neural network with the three-dimensional convolution kernel can capture cross-domain local semantic association features among the government feature matrix after training and correcting, the enterprise feature matrix after training and the college feature matrix after training and correcting, but the government feature matrix after training and correcting, the enterprise feature matrix after training and the college feature matrix after training and correcting have scale difference and unbalanced semantic density distribution in the expression of the feature source domain, so that the three-party association feature graph has more obvious inconsistency and instability of the overall feature distribution, thereby influencing the stability of classification training through a classifier.
Based on the above, when the applicant of the application performs classification training on the training three-party association feature map through the classifier, training optimization is performed on training three-party association feature vectors obtained by projecting the training three-party association feature map during each iteration, which is specifically expressed as follows: and training and optimizing the training three-party association feature vector obtained by the projection of the training three-party association feature map according to the following formula to obtain an optimized training three-party association feature vector.
Wherein, the formula is:
wherein,is the training three-party associated feature vector, +.>Is the training three-party associated feature vector +.>Characteristic value of>And->The training three-party associated feature vector>1-norm and 2-norm of +.>Is the training three-party associated feature vector +.>Length of (2), and->Is in combination with->Related weight superparameter +.>Is the characteristic value of the three-party associated characteristic vector of the optimization training.
Here, the feature vectors are correlated by the training three partiesStructural consistency and stability representation of the global feature distribution of (a) under rigid and non-rigid structures of absolute and spatial distances, respectively, such that the training three-party associated feature vector +.>Has a certain repeatability for local pattern changes to correlate feature vectors +.>When the classifier is used for classifying, robustness is provided for the scale and rotation change of the global feature distribution through the weight matrix of the classifier, and the stability of classification training is improved.
In summary, the collaborative innovation management system 200 according to the embodiment of the application is illustrated, which extracts global implicit associated feature distribution information of government parameters, enterprise parameters and college parameters in the collaborative innovation system through a deep neural network model, fuses feature information of the three by using a convolutional neural network with a three-dimensional convolution kernel, so as to mine dynamic implicit associated features of the three in a deeper level, and further performs cauchy weight probability correction on feature matrices of the three in the process, so that robustness to information loss is enhanced, and feature expression capability of a three-party associated feature map fusing feature information is improved. Thus, the efficiency of innovation in obstetric and research can be improved.
As described above, the collaborative innovation management system 200 according to the embodiment of the present application may be implemented in various terminal devices, such as a server of a collaborative innovation management algorithm, or the like. In one example, the collaborative innovation management system 200 according to an embodiment of the present application can be integrated into a terminal device as a software module and/or hardware module. For example, the collaborative innovation management system 200 can be a software module in the operating system of the terminal device or can be an application developed for the terminal device; of course, the collaborative innovation management system 200 can also be one of a number of hardware modules of the terminal device.
Alternatively, in another example, the collaborative innovation management system 200 and the terminal device may be separate devices, and the collaborative innovation management system 200 may be connected to the terminal device through a wired and/or wireless network and transmit interaction information in a agreed data format.
FIG. 3 illustrates a flow chart of a method of managing a collaborative innovation management system. As shown in fig. 3, a management method of a collaborative innovation management system according to an embodiment of the application includes the steps of: s110, acquiring a plurality of first parameters of a government, a plurality of second parameters of an enterprise and a plurality of third parameters of a university in a collaborative innovation system; s120, passing the first parameters of the government, the second parameters of the enterprise and the third parameters of the university through a context encoder comprising an embedded layer to obtain a plurality of government parameter feature vectors, a plurality of enterprise parameter feature vectors and a plurality of university parameter feature vectors; s130, two-dimensionally arranging the government parameter feature vectors into a first feature matrix, and then obtaining a government feature matrix through a first convolutional neural network; s140, two-dimensionally arranging the enterprise parameter feature vectors into a second feature matrix, and then obtaining an enterprise feature matrix through a second convolutional neural network; s150, two-dimensionally arranging the plurality of college parameter feature vectors into a third feature matrix, and then obtaining the college feature matrix through a third convolutional neural network; s160, performing Cauchy weight probability on the government feature matrix, the enterprise feature matrix and the college feature matrix to obtain a corrected government feature matrix, a corrected enterprise feature matrix and a corrected college feature matrix, wherein the Ke Xichong probability is performed based on the proportion between the feature values of each position in the feature matrix and the sum value of the feature values of all positions in the feature matrix; s170, constructing the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix into three-dimensional input tensors, and then obtaining a three-party association feature map through a fourth convolution neural network with a three-dimensional convolution kernel; and S180, the three-party association feature map is passed through a classifier to obtain a classification result, wherein the classification result is used for indicating whether abnormal collaborative innovation modes exist in governments, enterprises and universities in the collaborative innovation system.
Fig. 4 illustrates an architecture diagram of a management method of a collaborative innovation management system according to an embodiment of the application. As shown in fig. 4, in the network architecture of the management method of the collaborative innovation management system, first, a plurality of obtained first parameters (e.g., P1 as illustrated in fig. 4), a plurality of second parameters (e.g., P2 as illustrated in fig. 4) of the enterprise, and a plurality of third parameters (e.g., P3 as illustrated in fig. 4) of the university are respectively passed through a context encoder (e.g., E as illustrated in fig. 4) including an embedded layer to obtain a plurality of government parameter feature vectors (e.g., VF1 as illustrated in fig. 4), a plurality of enterprise parameter feature vectors (e.g., VF2 as illustrated in fig. 4), and a plurality of university parameter feature vectors (e.g., VF3 as illustrated in fig. 4); next, the plurality of government parameter feature vectors are two-dimensionally arranged into a first feature matrix (e.g., MF1 as illustrated in fig. 4) and then passed through a first convolutional neural network (e.g., CNN1 as illustrated in fig. 4) to obtain a government feature matrix (e.g., M1 as illustrated in fig. 4); then, the plurality of enterprise parameter feature vectors are two-dimensionally arranged into a second feature matrix (e.g., MF2 as illustrated in fig. 4) and then passed through a second convolutional neural network (e.g., CNN2 as illustrated in fig. 4) to obtain an enterprise feature matrix (e.g., M2 as illustrated in fig. 4); next, the plurality of college parameter feature vectors are two-dimensionally arranged into a third feature matrix (e.g., MF3 as illustrated in fig. 4) and then passed through a third convolutional neural network (e.g., CNN3 as illustrated in fig. 4) to obtain a college feature matrix (e.g., M3 as illustrated in fig. 4); then, cauchy re-probability is performed on the government feature matrix, the enterprise feature matrix, and the college feature matrix to obtain a corrected government feature matrix (e.g., MC1 as illustrated in fig. 4), a corrected enterprise feature matrix (e.g., MC2 as illustrated in fig. 4), and a corrected college feature matrix (e.g., MC3 as illustrated in fig. 4); next, the corrected government feature matrix, the corrected enterprise feature matrix, and the corrected college feature matrix are constructed as three-dimensional input tensors (e.g., DT as illustrated in fig. 4) followed by a fourth convolutional neural network with a three-dimensional convolutional kernel (e.g., CNN4 as illustrated in fig. 4) to obtain a three-way correlation feature map (e.g., F as illustrated in fig. 4); and finally, passing the three-party associated feature map through a classifier (e.g., as illustrated in fig. 4) to obtain a classification result, wherein the classification result is used for indicating whether the collaborative innovation mode of the government, the enterprise and the university in the collaborative innovation system is abnormal.
In summary, the management method of the collaborative innovation management system based on the embodiment of the application is clarified, global implicit associated feature distribution information of government parameters, enterprise parameters and college parameters in the collaborative innovation system is respectively extracted through a deep neural network model, the feature information of the three are fused by using a convolution neural network with a three-dimensional convolution kernel, so that dynamic implicit associated features of the three with deeper level are mined, and in the process, the feature matrix of the three is further subjected to Cauchy weight probabilistic correction, so that robustness to information loss is enhanced, and feature expression capability of a three-party associated feature map of the fused feature information is improved. Thus, the efficiency of innovation in obstetric and research can be improved.
The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not intended to be limited to the details disclosed herein as such.
The block diagrams of the devices, apparatuses, devices, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.
It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent to the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.
Claims (8)
1. A collaborative innovation management system, comprising:
the innovation parameter acquisition unit is used for acquiring a plurality of first parameters of a government, a plurality of second parameters of an enterprise and a plurality of third parameters of a university in the collaborative innovation system;
the parameter semantic coding unit is used for respectively passing the first parameters of the government, the second parameters of the enterprise and the third parameters of the university through a context encoder comprising an embedded layer to obtain a plurality of government parameter feature vectors, a plurality of enterprise parameter feature vectors and a plurality of university parameter feature vectors;
the government parameter association coding unit is used for two-dimensionally arranging the government parameter feature vectors into a first feature matrix and then obtaining the government feature matrix through a first convolutional neural network;
the enterprise parameter association coding unit is used for two-dimensionally arranging the plurality of enterprise parameter feature vectors into a second feature matrix and then obtaining an enterprise feature matrix through a second convolutional neural network;
The college parameter association coding unit is used for two-dimensionally arranging the plurality of college parameter feature vectors into a third feature matrix and then obtaining the college feature matrix through a third convolutional neural network;
the characteristic distribution correction unit is used for carrying out cauchy weight probability on the government characteristic matrix, the enterprise characteristic matrix and the college characteristic matrix to obtain a corrected government characteristic matrix, a corrected enterprise characteristic matrix and a corrected college characteristic matrix, wherein the Ke Xichong probability is carried out based on the proportion between the characteristic value of each position in the characteristic matrix and the sum value of the characteristic values of all positions in the characteristic matrix;
the three-party association coding unit is used for constructing the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix into three-dimensional input tensors and then obtaining a three-party association feature map through a fourth convolution neural network with a three-dimensional convolution kernel; and
the innovation mode evaluation result generation unit is used for enabling the three-party association feature graphs to pass through a classifier to obtain classification results, wherein the classification results are used for indicating whether abnormal collaborative innovation modes exist in governments, enterprises and universities in the collaborative innovation system;
Wherein, still include: training means for training the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the fourth convolutional neural network with three-dimensional convolutional kernels, and the classifier;
wherein, training module includes:
the training innovation parameter acquisition unit is used for acquiring training data, wherein the training data comprises a plurality of training first parameters of a government, a plurality of training second parameters of an enterprise and a plurality of training third parameters of a college in a collaborative innovation system;
the training parameter semantic coding unit is used for respectively enabling a plurality of training first parameters of the government, a plurality of training second parameters of the enterprise and a plurality of training third parameters of the college to pass through a context encoder comprising an embedded layer to obtain a plurality of training government parameter feature vectors, a plurality of training enterprise parameter feature vectors and a plurality of training college parameter feature vectors;
the training government parameter association coding unit is used for two-dimensionally arranging the plurality of training government parameter feature vectors into a training first feature matrix and then obtaining the training government feature matrix through the first convolutional neural network;
The training enterprise parameter association coding unit is used for performing two-dimensional arrangement on the plurality of training enterprise parameter feature vectors to obtain a training enterprise feature matrix through the second convolutional neural network after training the second feature matrix;
the training college parameter association coding unit is used for performing two-dimensional arrangement on the plurality of training college parameter feature vectors to form a training third feature matrix, and then obtaining the training college feature matrix through the third convolutional neural network;
the training feature distribution correction unit is used for carrying out cauchy weight probability on the training government feature matrix, the training enterprise feature matrix and the training college feature matrix to obtain a training corrected government feature matrix, a training corrected enterprise feature matrix and a training corrected college feature matrix;
the training three-party association coding unit is used for constructing the training corrected government feature matrix, the training corrected enterprise feature matrix and the training corrected college feature matrix into training three-dimensional input tensor and then obtaining a training three-party association feature map through the fourth convolution neural network with the three-dimensional convolution kernel; and
the classification loss calculation unit is used for enabling the training three-party association feature images to pass through the classifier to obtain classification loss function values;
The back propagation training unit is used for training the first convolutional neural network, the second convolutional neural network, the third convolutional neural network, the fourth convolutional neural network with a three-dimensional convolutional kernel and the classifier by propagating the classification loss function value in a gradient descending direction, wherein training optimization is carried out on training three-party association feature vectors obtained by projection of the training three-party association feature map during each iteration;
and performing training optimization on the training three-party associated feature vector obtained by the training three-party associated feature map projection during each iteration, wherein the training optimization comprises the following steps: training and optimizing the training three-party association feature vector obtained by the projection of the training three-party association feature map by using the following formula to obtain an optimized training three-party association feature vector;
wherein, the formula is:
wherein V is the training three-party associated feature vector, V i Is the feature value of the training three-party associated feature vector V, I 1 And|| | 2 Is the 1-norm and 2-norm of the training three-party associated feature vector V, L is the length of the training three-party associated feature vector V, and alpha is the sum of V i Related weight superparameter, v' i Is the characteristic value of the three-party associated characteristic vector of the optimization training.
2. The collaborative innovation management system of claim 1, wherein the parameter semantic coding unit comprises:
an input vector construction subunit for constructing a plurality of first parameters of the government, a plurality of second parameters of the enterprise, and a plurality of third parameters of the university as input vectors, respectively, to obtain a sequence of input vectors;
an embedding subunit for converting each input vector in the sequence of input vectors into an embedded vector using an embedding layer of the context encoder including an embedding layer to obtain a sequence of embedded vectors; and
a context semantic encoding subunit configured to perform global-based context semantic encoding on the sequence of embedded vectors using a translator of the context encoder including an embedded layer to obtain the plurality of government parameter feature vectors, the plurality of enterprise parameter feature vectors, and the plurality of college parameter feature vectors.
3. The collaborative innovation management system of claim 2, wherein the government parameter-associated encoding unit is further configured to use the layers of the first convolutional neural network to separately perform, during forward transfer of the layers, input data:
Carrying out convolution processing on input data to obtain a convolution characteristic diagram;
carrying out local channel dimension-based mean pooling on the convolution feature map to obtain a pooled feature map; and
non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map;
wherein the output of the last layer of the first convolutional neural network is the government feature matrix, and the input of the first layer of the first convolutional neural network is the first feature matrix.
4. The collaborative innovation management system of claim 3, wherein the college parameter association encoding unit is further configured to: and respectively carrying out convolution processing, local channel dimension average pooling processing and nonlinear activation processing on input data in the forward transfer process of the layers by using each layer of the third convolutional neural network so as to output the college feature matrix by the last layer of the third convolutional neural network.
5. The collaborative innovation management system of claim 4, wherein the feature distribution correction unit is further configured to: performing cauchy re-probability on the government feature matrix, the enterprise feature matrix and the college feature matrix respectively according to the following formulas to obtain the corrected government feature matrix, the corrected enterprise feature matrix and the corrected college feature matrix;
Wherein, the formula is:
wherein m is i,j A feature value for each location of each of the government feature matrix, the enterprise feature matrix, and the college feature matrix, andthe method comprises the steps of summing eigenvalues of all positions of a feature matrix, exp (·) represents an exponential operation of the matrix, the exponential operation of the matrix represents a natural exponential function value calculated by exponentiating the eigenvalues of all positions in the matrix, and the division of the feature matrix by a parameter represents the division of the eigenvalues of all positions in the feature matrix by the parameter.
6. The collaborative innovation management system of claim 5, wherein the three-party association encoding unit is further configured to: the fourth convolutional neural network with the three-dimensional convolutional kernel performs the respective processing on the input data in the forward transfer of the layer:
performing three-dimensional convolution processing on the input data based on the three-dimensional convolution check to obtain a convolution characteristic diagram;
carrying out mean pooling treatment based on a local feature matrix on the convolution feature map to obtain a pooled feature map; and
non-linear activation is carried out on the pooled feature map so as to obtain an activated feature map;
the output of the last layer of the fourth convolutional neural network is the three-party association feature map, and the input of the first layer of the fourth convolutional neural network is the three-dimensional input tensor.
7. The collaborative innovation management system of claim 6, wherein the innovation pattern evaluation result generation unit is further configured to: the classifier processes the three-party association feature map to generate a classification result according to the following formula;
wherein, the formula is: softmax { (W) n ,B n ):…:(W 1 ,B 1 ) Project (F), where Project (F) represents projecting the three-party associated feature map into three-party associated feature vectors, W 1 To W n Weight matrix for all the connection layers of each layer, B 1 To B n Representing the bias matrix for each fully connected layer.
8. The collaborative innovation management system of claim 7, wherein the first plurality of government parameters includes a total cost of government participation in collaborative innovation, a benefit obtained by government selection "participate" policy, a proportion of benefit obtained by government selection "do not participate" policy to benefit obtained by government selection "participate" policy, and a willingness of government selection to participate in collaborative innovation; the second parameters of the enterprise include a reduction amount of total cost input by the enterprise and the university in the collaborative innovation process by a preferential policy provided by the government, initial benefits before the enterprise performs the collaborative innovation, benefits obtained by independent research and development of the enterprise, punishments paid to the enterprise when the enterprise breaks down, willingness of the enterprise to select the collaborative innovation, additional benefits brought to the enterprise when the enterprise and the university select the collaborative innovation, total cost of the enterprise and the university participating in the collaborative innovation and cost sharing proportionality coefficients of the enterprise; and, the plurality of third parameters of the university include a governmental provided benefit policy such that the enterprise and the university are reduced in total cost invested in the collaborative innovation process, an initial benefit before the collaborative innovation is performed by the university, a benefit obtained by the independent development of the university, a penalty paid to the enterprise when the university breaks, a willingness of the university to choose to perform the collaborative innovation, an additional benefit brought to the university when the enterprise and the university both choose to perform the collaborative innovation, a total cost of the enterprise and the university to participate in the collaborative innovation, a fund support given by the government to the university with the active parameter collaborative innovation, and a cost allocation scaling factor of the university.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311685364.3A CN117391646B (en) | 2023-12-11 | 2023-12-11 | Collaborative innovation management system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311685364.3A CN117391646B (en) | 2023-12-11 | 2023-12-11 | Collaborative innovation management system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117391646A CN117391646A (en) | 2024-01-12 |
| CN117391646B true CN117391646B (en) | 2024-03-22 |
Family
ID=89470601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311685364.3A Active CN117391646B (en) | 2023-12-11 | 2023-12-11 | Collaborative innovation management system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117391646B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108734621A (en) * | 2018-03-23 | 2018-11-02 | 江苏科技大学 | A kind of political affairs production, teaching & research collaborative innovation tripartite's evolutionary Game Analysis method |
| CN108846340A (en) * | 2018-06-05 | 2018-11-20 | 腾讯科技（深圳）有限公司 | Face identification method, device and disaggregated model training method, device, storage medium and computer equipment |
| CN114911958A (en) * | 2022-06-09 | 2022-08-16 | 电子科技大学 | Semantic preference-based rapid image retrieval method |
| CN115147655A (en) * | 2022-07-12 | 2022-10-04 | 温州宁酷科技有限公司 | Oil gas gathering and transportation monitoring system and method thereof |
| CN115796621A (en) * | 2022-12-23 | 2023-03-14 | 国网电力科学研究院有限公司 | Comprehensive energy system development analysis method and system based on evolutionary game theory |
| CN116307841A (en) * | 2023-02-14 | 2023-06-23 | 河海大学 | Construction method of college evaluation system integrating multiple indexes |
-
2023
- 2023-12-11 CN CN202311685364.3A patent/CN117391646B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108734621A (en) * | 2018-03-23 | 2018-11-02 | 江苏科技大学 | A kind of political affairs production, teaching & research collaborative innovation tripartite's evolutionary Game Analysis method |
| CN108846340A (en) * | 2018-06-05 | 2018-11-20 | 腾讯科技（深圳）有限公司 | Face identification method, device and disaggregated model training method, device, storage medium and computer equipment |
| CN114911958A (en) * | 2022-06-09 | 2022-08-16 | 电子科技大学 | Semantic preference-based rapid image retrieval method |
| CN115147655A (en) * | 2022-07-12 | 2022-10-04 | 温州宁酷科技有限公司 | Oil gas gathering and transportation monitoring system and method thereof |
| CN115796621A (en) * | 2022-12-23 | 2023-03-14 | 国网电力科学研究院有限公司 | Comprehensive energy system development analysis method and system based on evolutionary game theory |
| CN116307841A (en) * | 2023-02-14 | 2023-06-23 | 河海大学 | Construction method of college evaluation system integrating multiple indexes |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117391646A (en) | 2024-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Shao et al. | Branchy-GNN: A device-edge co-inference framework for efficient point cloud processing | |
| Vershynin | Four lectures on probabilistic methods for data science | |
| US11842250B2 (en) | Quantum error correction decoding system and method, fault-tolerant quantum error correction system, and chip | |
| CN116095089B (en) | Remote sensing satellite data processing method and system | |
| CN104657434B (en) | A kind of social network structure construction method | |
| CN110363068A (en) | A kind of high-resolution pedestrian image generation method based on multiple dimensioned circulation production confrontation network | |
| CN112446489A (en) | Dynamic network embedded link prediction method based on variational self-encoder | |
| Wang et al. | A research on the optimal design of BP neural network based on improved GEP | |
| Bi et al. | Knowledge transfer for out-of-knowledge-base entities: Improving graph-neural-network-based embedding using convolutional layers | |
| Tanaka et al. | A faithful distributed implementation of dual decomposition and average consensus algorithms | |
| CN117391646B (en) | Collaborative innovation management system | |
| CN116030537B (en) | Three-dimensional human body posture estimation method based on multi-branch attention-seeking convolution | |
| CN115527052A (en) | Multi-view clustering method based on contrast prediction | |
| CN116228014A (en) | DC power distribution network benefit evaluation system and method for infrastructure network access | |
| Aketi et al. | Neighborhood gradient clustering: An efficient decentralized learning method for non-iid data distributions | |
| CN115861664A (en) | Feature matching method and system based on local feature fusion and self-attention mechanism | |
| CN115660882A (en) | Method for predicting user-to-user relationship in social network and multi-head mixed aggregation graph convolutional network | |
| Wang et al. | Heterogeneous defect prediction algorithm combined with federated sparse compression | |
| CN115345298A (en) | Method and device for jointly training models | |
| Tziotis et al. | Second order optimality in decentralized non-convex optimization via perturbed gradient tracking | |
| CN115204171A (en) | Document-level event extraction method and system based on hypergraph neural network | |
| CN114880527A (en) | Multi-modal knowledge graph representation method based on multi-prediction task | |
| CN115130568A (en) | Longitudinal federated Softmax regression method and system supporting multiple parties | |
| CN114662148A (en) | Multi-party combined training method and device for protecting privacy | |
| CN114139721A (en) | Distributed learning ciphertext calculation efficiency improving method based on homomorphic encryption |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |