30 results about "Functional similarity" patented technology

Disclosed is a cloud-native application programming interface (API) recommendation method fusing data augmentation and contrastive learning. Service information is included on the basis of a service information double-graph structure, and a mutual attention mechanism is designed to compute an importance degree of each layer of information. A data optimization method for sequence information based on functional similarity and a computation method for similarity between services based on two parts of information are provided; on this basis, data of a service invocation sequence is augmented with the idea of contrastive learning to form an augmented sequence pair; a computational contrastive loss function is combined with a pair-wise recommendation loss function to optimize an overall model, thereby improving the effect of a service recommendation model; and according to a feature embedding representation result of a service, pair-wise recommendation scores are computed to complete service recommendation.

The invention discloses a network security vulnerability identification and repair method and system, a medium and a program product, and the method comprises the steps: segmenting a system behavior sequence into a plurality of subsequences; constructing a behavior dependency graph; calculating functional similarity, and dividing the functional similarity into node clusters; calculating a behavior consistency score, and determining a potential vulnerability region; determining call chain information; constructing a vulnerability propagation path, and calculating criticality; dividing first-level hazard vulnerabilities, second-level hazard vulnerabilities and third-level hazard vulnerabilities according to the criticality; determining a repair priority, and generating a vulnerability report; according to the vulnerability types, the repair priorities and the propagation path information, corresponding repair strategies are sequentially matched from a preset repair template library according to the sequence of the repair priorities; and performing hierarchical repair on the application code based on the repair strategy obtained by matching to obtain a repair code, and deploying the repair code to the target system. According to the invention, the accuracy and efficiency of identifying and repairing network security vulnerabilities are improved.

The invention discloses a method for analyzing software code similarity and an analysis system, and relates to the field of software engineering and machine learning. The method comprises the steps of establishing a code snippet database, extracting multi-dimensional features of code snippets, executing similarity analysis of the code snippets, and establishing a similarity analysis result; performing context backtracking analysis on the code snippets in the database by using an adaptive analysis network, and establishing a context similarity result; performing static and dynamic fusion analysis on the database, establishing a joint function similarity, and establishing a function similarity result; code version change analysis is executed, and a version similarity result is established; and performing code similarity retrieval management according to the four analysis results. The technical problem that the code similarity analysis efficiency and precision are low due to the fact that the code snippets are high in complexity in multi-dimensional feature analysis is solved, and the code similarity analysis accuracy and efficiency are improved in the mode that multi-dimensional analysis and self-adaptive backtracking analysis are combined.

The invention designs a miRNA-disease association prediction method based on an adaptive content guided fusion mechanism and multi-scale gated convolution. Firstly, miRNA function similarity, disease semantic similarity and Gaussian interaction spectrum kernel similarity of the miRNA function similarity and the disease semantic similarity are calculated, the weight of each similarity matrix is adjusted through a self-adaptive content guide fusion mechanism, and similarity information contributing to a prediction task is automatically concerned. The method comprises the following steps: firstly, carrying out feature extraction through multi-scale gating convolution by combining Laplacian features and Weisfeier-Lehman absolute role coding features, and finally, connecting learned feature representations of miRNA and diseases, and sending the feature representations into an XGBoost classifier for training and predicting potential association of miRNA and diseases. The working mechanism of miRNA in diseases is deeply explored, practical application value is achieved for clinical diagnosis and prevention, and meanwhile discovery of a novel therapy is promoted.

The invention provides an LncRNA-disease association prediction method based on comparative learning. Comprising the following steps: S1, preparing an LncRNA-disease incidence matrix, calculating functional similarity, cosine similarity and Gaussian interaction spectrum kernel similarity of lncRNA, and calculating semantic similarity, cosine similarity and Gaussian interaction spectrum kernel similarity of diseases; s2, constructing an LncRNA similarity graph, a disease similarity graph and an LncRNA disease association graph; s3, obtaining an LncRNA feature matrix, an LncRNA auxiliary feature matrix, a disease feature matrix and a disease auxiliary feature matrix; s4, obtaining an LncRNA personalized fusion feature and a disease personalized fusion feature; and S5, carrying out cross-view comparative learning on the personalized fusion features of the LncRNA and the personalized fusion features of the diseases, and predicting potential LncRNA and disease association. According to the method, information fused into bioinformatics is considered, so that the method has certain value for predicting the potential correlation between the LncRNA and the diseases, improving the accuracy of disease diagnosis and treatment and promoting the development of drug research and development.

This invention relates to the field of Internet of Things (IoT) technology and is a cascading threat detection method for automated rules in IoT. It utilizes a large language model to parse rule text, extract semantic elements, and construct a semantic representation model. A semantic alignment model is built to measure the functional similarity of different rules in a rule set. A TAP rule heterogeneous graph is constructed, establishing explicit and implicit relationships respectively. A dual-attention context encoder is constructed to obtain rule embedding representations containing semantic dependency information. A global relation attention mechanism is introduced to measure the importance of explicit and implicit relation spaces under different node types. The invention determines whether cascading threat paths exist in the rule set and generates threat detection results. This invention can automatically parse rule semantics, construct explicit and implicit dual-relation graph models, and utilize multi-layer graph neural networks to achieve multi-hop inference, thereby comprehensively detecting potential rule-cascading threats.

The application relates to a cell infiltration inference method and system fusing GO function annotation and PPI network information, and the method comprises the following steps: collecting gene expression data, GO function annotation data and PPI network data; constructing a cell-cell function correlation network and a cell-cell physical interaction network respectively; performing weighted fusion processing on the two networks to obtain a comprehensive cell relationship network; calculating a final cell infiltration score through a restart walk algorithm, and inferring the infiltration degree in a tumor microenvironment according to the final cell infiltration score. The application innovatively fuses GO function annotation information and PPI network data, comprehensively considers the functional similarity and physical or signal interaction between cells, enables the model to understand cell synergy from the biological pathway level and analyze cell direct interaction from the protein interaction level, avoids one-sidedness of a single perspective, and provides a more stereoscopic cognitive framework for tumor microenvironment analysis.

The invention discloses a universal digital project function similarity comparison method, which is completed through S1 digital project function identification, S2 function objectification processing and S3 function similarity calculation, and comprises the following steps: S1, carrying out identification and function division on a project, constructing a function list and carrying out hierarchical description on the function list; the processing process in the step S2 is to convert the function list into a function object with a unique code, a function title object and a function description object, and comprises S2-1 function object processing, S2-2 function title object processing and S2-3 function description object processing; in the step S3, similarity calculation is carried out from bottom to top, function title similarity, function description similarity and function similarity are obtained respectively, and finally similarity values of all function objects are obtained; according to the invention, the problem of pain points of digital project repetitive review is solved, an intelligent project review system is constructed, and a digital project repetitive review function is realized.

The invention relates to a cell infiltration inference method and system fusing GO function annotation and PPI network information. The method comprises the following steps: collecting gene expression data, GO function annotation data and PPI network data; respectively constructing a cell * cell function association network and a cell * cell physical interaction network; carrying out weighted fusion processing on the two to obtain a comprehensive cell relation network; and calculating a final cell infiltration fraction through a restart migration algorithm, and deducing the infiltration degree in the tumor microenvironment according to the final cell infiltration fraction. According to the method, GO function annotation information and PPI network data are creatively fused, functional similarity and physical or signal interaction between cells are comprehensively considered, the model can understand cell synergy from the biological pathway level and can analyze cell direct interaction from the protein interaction level, one-sidedness of a single view angle is avoided, and the method has the advantages of being simple in structure and convenient to operate. And a more three-dimensional cognitive framework is provided for tumor microenvironment analysis.

The invention provides a physiological signal channel aggregation method and system for a graph neural network. The method comprises the steps of obtaining channel node features and an adjacent matrix of an upstream task; calculating the similarity between channel nodes, initializing a sorting matrix, and carrying out row and column normalization on the sorting matrix; sorting the channel node features and the adjacent matrixes by using a sorting matrix; the method comprises the following steps: grouping channels, constructing mapping, and converting a sequenced adjacent matrix into a mapping matrix; aggregating the sorted channel node features and the mapping matrix; and constructing a regional aggregation loss function to constrain the adjacent matrix. The technical problems that universality is poor, functional similarity changes between channels are difficult to dynamically capture, knowledge in the preset field is relied on, and capture features are single are solved.

The present application discloses a method for annotating arbitrary code snippets and a retrieval system thereof, which relates to the field of machine learning. The method comprises: establishing a code snippet database, activating a multidimensional feature annotation model, performing code snippet annotation, and establishing a first annotation result; calling an adaptive annotation network, utilizing the adaptive annotation network to perform contextual backtracking analysis on the code snippets in the database, and establishing a second annotation result; performing static and dynamic fusion analysis on the database, establishing joint functional similarity, and establishing a third annotation result; performing code version change annotation, and establishing a fourth annotation result; and performing code annotation retrieval management based on the four annotation results. The method solves the technical problem of low management efficiency and retrieval accuracy of code snippets due to the high complexity of code snippets in multidimensional feature annotation, and achieves the technical effect of improving the annotation accuracy and retrieval efficiency of code snippets by combining multidimensional annotation with adaptive backtracking analysis.

The present invention discloses a method for constructing a spatiotemporal knowledge graph. For each device in the smart home space, the specification document is respectively subjected to image text matching and text structured extraction to construct a specification knowledge graph for each device. Then, the information contained in the device specification knowledge graph is used to filter out valid log information pairs in the smart home space, and the smart home spatiotemporal knowledge graph is constructed according to the joint functional similarity between the devices contained in the log pairs. The present invention utilizes the device information contained in the specification document and the timing information and user behavior patterns contained in the user behavior log, and combines the knowledge graph and the large language model to understand and construct a comprehensive, dynamic and time-space evolving smart home domain knowledge graph, which can not only optimize the collaborative work between devices, but also significantly improve the intelligence level of the system, so that users can more conveniently manage and control the home environment, and realize more personalized and intelligent home services.

The application belongs to the technical field of picture processing, and particularly relates to a picture generation method based on mutation data, a generation system and a cancer metastasis prediction method. Step one: constructing a pathway image by using the functional similarity of pathways on a gene interaction network; step two: constructing a patient characteristic image; the prediction method further comprises: training a prediction model by using the constructed patient characteristic image, and predicting the metastasis of mutation data by using the trained prediction model. The application aims to solve the problems of lacking high-performance generation of patient characteristic pictures based on single nucleotide variation and the technical problems of cancer metastasis prediction and determination.

The invention discloses a method for predicting the association between lncRNA and disease based on a collaborative matrix decomposition of high-order perturbations and flexible ternary regularization, which belongs to the technical field of matrix decomposition and biological gene combination. The method comprises the following specific steps: obtaining the association matrix Z of lncRNA and disease; calculating the functional similarity matrix R of lncRNA based on the known association matrix Z; s and the semantic similarity matrix D of the disease s ; The multi-order correlation matrix calculated based on the known correlation matrix is ​​added to the known correlation matrix as a high-order perturbation; the global nonlinear structural similarity information of the similarity matrix is ​​obtained through similarity learning of flexible ternary regular terms; the objective function is optimized and iterated based on the collaborative matrix decomposition method to obtain the final prediction score matrix, so that the final prediction of lncRNA-disease association is more accurate.

The invention belongs to the technical field of enzyme catalytic conversion number prediction, and discloses an enzyme catalytic conversion number prediction method and system based on function annotation and a hierarchical structure, and the prediction method comprises the steps: extracting protein unique identifier information based on a protein resource database, analyzing protein dynamic information and gene ontology functions according to the unique identifier information, and obtaining a prediction result; obtaining a relation table of the gene ontology and the enzymatic conversion coefficient; combining the hierarchical structure of the gene ontology with the relation table, capturing the mutual relation between the gene and the gene product, and constructing a hierarchical total tree between the gene ontology and the enzymatic conversion coefficient according to a relation capturing result; and extracting target gene ontology information according to the gene code of the target object, and matching an enzymatic conversion coefficient corresponding to the target gene ontology information from the hierarchical total tree as an enzymatic conversion number prediction result. According to the invention, gene and protein function annotations are provided by using the gene ontology, so that the function similarity of enzymes can be measured based on the gene ontology, and the enzymatic conversion value of unknown enzymes is speculated.

The present invention relates to the field of bioinformatics association prediction technology, and specifically to an encoder-based gradient boosting machine miRNA-disease association prediction method. The method comprises: utilizing multi-source biological and medical information to obtain a miRNA-disease association adjacency matrix, a miRNA functional similarity matrix, and a disease semantic similarity matrix; concatenating the resulting integrated miRNA-disease similarity matrix with the miRNA-disease association adjacency matrix to obtain a more informative miRNA-disease association feature vector; utilizing an autoencoder to extract key features of the combined miRNA feature vector and disease feature vector, and utilizing a lightweight gradient boosting machine classifier to predict potential associations between miRNAs and diseases. The present invention achieves high prediction accuracy with low time and economic costs, aims to uncover potential miRNA-disease associations, and can aid in the study of the pathogenesis of complex diseases.

The present disclosure provides an instruction matching method and apparatus, a wearable device, and a storage medium, which belong to the field of instruction matching technology. The method includes: determining an instruction list based on a function type and an interactive action type. Determining a function priority sequence corresponding to a function type and an action priority sequence corresponding to an interactive action type based on historical usage data. Determining the functional similarity of multiple functions in a function priority sequence based on the function type, and determining a function-action matching relationship based on the function priority sequence, the action priority sequence, and the functional similarity. Determining an instruction matching relationship based on the function-action matching relationship and the instruction list. The instruction matching relationship includes an action-instruction-function matching relationship. The instruction matching method and apparatus, wearable device, and storage medium provided by the present disclosure can solve the problem that the prior art lacks an effective instruction matching method, and improve the user experience.

Quantification and Detection of Immune Landscape Changes An immune response to an event is detected and quantified by accessing sets of immune cell sequences taken from a sample before and after an event, detecting the immune cell sequences showing a significant change between before and after the event, clustering the immune cell sequences so as to cluster together cell sequences having functional similarity, selecting the clusters containing at least one immune cell sequence showing a significant change and calculating an immune response score based on the selected clusters.

A circRNA-disease association prediction method based on graph convolutional neural network and Node2vec, the steps are as follows: obtain circRNA-disease association matrix; calculate circRNA functional similarity, circRNA Gaussian interaction spectrum kernel similarity, disease Gaussian interaction spectrum kernel similarity, disease semantic similarity, construct circRNA integrated similarity and disease integrated similarity, and generate circRNA-disease heterogeneous graph; sparse autoencoder extracts features of circRNA (disease) integrated similarity and transforms it into a 64-dimensional feature vector, and fuses the circRNA feature vector and the disease feature vector into the final circRNA-disease feature vector; graph convolutional neural network extracts local structural information of nodes from circRNA-disease heterogeneous graph; Node2vec method extracts global structural information of nodes from circRNA-disease heterogeneous graph; the node information obtained in the first two steps is sent to random forest classifier to predict potential circRNA-disease association. Predicting disease-related circRNA by computational method saves time, helps to clarify the pathogenesis of the disease, and finds effective treatment plans.

The invention discloses a multifunctional peptide design method and application thereof, the method comprises sequence allosteric operation, independent of splicing of different functional fragments, and capability of improving a single-function polypeptide to enable the single-function polypeptide to have multiple functions, and the specific method is as follows: on the tendency of a first functional template sequence, individual amino acid mutation is carried out to enable the first functional template sequence to be converted into second functional similarity; the polypeptide is endowed with multifunctionality and can be simplified and understood that a characteristic functional sequence and a characteristic polypeptide structure respectively perform different functions, so that the effect that a single sequence has multiple functions is achieved.

The present invention belongs to the field of model security technology, and discloses a model security assessment method and system for hard label no-data setting, the method comprising step S1, sending the target model classification sample into the denoising diffusion model to generate proxy data, step S2, sending the noise data into the generator, training the generator to generate query samples, step S3, training the substitute model through class diversity loss to perform functional fitting on the target model, step S4, calculating the functional similarity of the substitute model, and evaluating the security of the target model. The present invention can accurately assess the security of the black box model deployed online.

The invention discloses an AI-assisted test vector analysis and optimization method and system, and relates to the technical field of integrated circuit design automation. The method comprises the following steps: a feature extraction step: obtaining a test vector set and extracting multi-dimensional features including fault coverage fingerprints and a scan chain load mode, and constructing a feature matrix; an intelligent clustering step: clustering the test vector set by using an unsupervised clustering algorithm, and dividing vectors with similar functions into the same vector cluster; a representative vector selection step: selecting representative vectors in each vector cluster based on a fault coverage integrity principle to form an optimized vector set; and an iterative optimization step: verifying the optimized vector set, and if a preset condition is not met, dynamically adjusting the clustering parameter and repeating the previous steps. According to the method, deep compression of the test vector is realized through global function similarity analysis, the test time and the test cost can be remarkably reduced, and the method has the advantages of full-automatic post-processing, support of multi-objective optimization and the like.

The invention relates to a city charging prediction method and device based on a multi-semantic topological graph and a medium, and the method comprises the steps: dividing a target city region into a plurality of space nodes, obtaining the space information and semantic information of each space node, and constructing a geographic adjacency matrix and a plurality of single semantic similarity matrixes; constructing a long-range connection candidate set based on the single semantic similarity matrix; obtaining a scene demand, and performing sparsification on the long-range connection candidate set based on the scene demand to obtain a long-range shortcut matrix; constructing a multi-semantic topological graph of the target area based on the geographic adjacency matrix and the long-range shortcut matrix; historical city charging data is acquired, and the future city charging demand is predicted by using the space-time diagram neural network based on the multi-semantic topological graph and the historical city charging data. Compared with the prior art, the method has the advantages that a traditional topological graph is optimized by fusing multiple spatial semantics such as geographical adjacency, functional similarity, travel modes and behavior modes, and therefore more accurate charging prediction is achieved.

Disclosed is a cloud-native application programming interface (API) recommendation method fusing data augmentation and contrastive learning. Service information is included on the basis of a service information double-graph structure, and a mutual attention mechanism is designed to compute an importance degree of each layer of information. A data optimization method for sequence information based on functional similarity and a computation method for similarity between services based on two parts of information are provided; on this basis, data of a service invocation sequence is augmented with the idea of contrastive learning to form an augmented sequence pair; a computational contrastive loss function is combined with a pair-wise recommendation loss function to optimize an overall model, thereby improving the effect of a service recommendation model; and according to a feature embedding representation result of a service, pair-wise recommendation scores are computed to complete service recommendation.

The invention relates to the technical field of machine learning, and discloses an miRNA subcellular localization prediction method and system based on hypergraph and contrast learning. The method comprises the following steps: extracting miRNA-disease associated characteristics and miRNA-mRNA associated network characteristics of a miRNA function similarity network, a miRNA-mRNA associated network and a miRNA-disease associated network by adopting a constructed hypergraph; carrying out comparative learning fusion processing on the miRNA sequence features, the miRNA-disease associated features and the miRNA-mRNA associated network features to obtain comparative learning fusion features; and establishing a miRNA subcellular localization prediction model according to the contrast learning fusion features, and carrying out localization based on the miRNA subcellular localization prediction model. The correlation between miRNA and mRNA and diseases is fully utilized, and positive and negative samples are compared by contrast learning, so that the accuracy of miRNA subcellular localization prediction is effectively improved.

The present invention provides a lncRNA-disease association prediction method based on a weighted kernel canonical regularization algorithm, comprising the following steps: S1: obtaining the lncRNA-disease adjacency matrix LD; S2: calculating the lncRNA expression similarity LS exp , lncRNA functional similarity LS fun , lncRNA Gaussian similarity LS gau , lncRNA linear neighborhood similarity LS lin , disease semantic similarity DS sem , disease Gaussian similarity DS gau , disease linear neighborhood similarity DS lin S3: The k-nearest-neighbor centered kernel alignment algorithm is used to integrate the similarities of lncRNAs and diseases into the same space. S4: A heterogeneous matrix is ​​constructed using the lncRNA-disease association matrix and the optimal similarity kernel matrix of lncRNAs and diseases as the target matrix for matrix completion. S5: The heterogeneous matrix is ​​input into a weighted kernel canonical regularization model for completion, ultimately obtaining the predicted lncRNA-disease association. This method utilizes the k-nearest-neighbor centered kernel alignment algorithm to integrate similarity information for auxiliary prediction and constructs a weighted kernel canonical regularization model to complete the heterogeneous matrix, achieving more accurate lncRNA-disease association prediction.

The present invention is applicable to the technical field of lattice structures, and provides a method for designing a bionic honeycomb lattice structure, the method comprising: starting from the characteristics of the honeycomb structure, extracting the main structural elements and structural mechanisms through similarity analysis and extraction of the honeycomb structure, confirming the selected elements from the perspectives of structural similarity, functional similarity and boundary similarity through a similarity analysis method according to the bionics principle; extracting structural features according to the characteristics of the honeycomb structure and fusing them with the concave triangle structure to establish a three-dimensional macroscopic preliminary bionic honeycomb single cell structure, and establishing the range of each element and its size for structural construction and dimensional parameter optimization; the content of the present invention: the lightweight degree and energy absorption performance of parts can be improved.

The invention relates to a concept innovation-oriented analogy knowledge clustering method. The method comprises the following steps of: 1, determining a function similarity key characteristic index system of knowledge clustering; 2, performing knowledge retrieval according to the key function indexes; 3, calculating the measurement of the functional similarity between the clustering knowledge according to a cosine law method, and setting a threshold value according to a calculation result; 4, calculating a technical difference degree parameter of each piece of knowledge to be clustered; and step 5, completing clustering of knowledge of various ratios through calculation results of the two clustering indexes, namely the function similarity and the technical difference. According to the cross-domain knowledge clustering method, efficient cross-domain knowledge clustering can be effectively achieved, analogy knowledge in different fields can be effectively processed through the clustering method combining the two dimensions of function similarity and technical difference, efficient clustering of cross-domain knowledge is achieved, and generation of higher-quality subversive innovation concepts is facilitated.

The invention provides a miRNA-lncRNA interaction prediction system based on multi-view projection fusion and truncated matrix decomposition, and belongs to the crossing field of bioinformatics and machine learning. The method comprises the following steps: firstly, constructing a comprehensive miRNA-lncRNA similarity map by comprehensively utilizing sequence similarity, functional similarity, expression similarity and Gaussian kernel similarity of miRNA / lncRNA, and constructing a corresponding network structure based on different similarities and known interaction information; secondly, a multi-view projection fusion method is provided, similar network topology structures from multiple view angles and original interaction matrix information are fully fused, and abundant topology features in the biological network are mined while the integrity of the interaction matrix structure is maintained. Finally, a truncated matrix decomposition technology is introduced, matrix dimension reduction is achieved through the number of truncated singular values, key feature information is reserved, and therefore prediction efficiency and precision are improved.

The invention relates to the technical field of Internet of Things, in particular to an Internet of Things automation rule-oriented cascade threat detection method, which comprises the following steps of: analyzing a rule text by using a large language model, extracting semantic elements and constructing a semantic representation model; constructing a semantic alignment model, and measuring functional similarities of different rules in the rule set; constructing a TAP rule heterogeneous graph, and respectively establishing an explicit relationship and an implicit relationship; constructing a double-attention context encoder to obtain a rule embedding representation containing semantic dependency information; introducing a global relation attention mechanism, and measuring the importance of an explicit relation space and an implicit relation space under different node types; and judging whether a cascade threat path exists in the rule set, and generating a threat detection result. According to the method, rule semantics can be automatically analyzed, the explicit and implicit double-relation graph model is constructed, and multi-hop reasoning is realized by using the multi-layer graph neural network, so that potential rule cascade threats are comprehensively detected.