19 results about "Relational structure" patented technology

The invention discloses a file information association analysis method based on a heterogeneous knowledge graph, which comprises the following steps: collecting and preprocessing multi-modal heterogeneous file information to generate a standardized input data set; establishing a heterogeneous knowledge graph based on the standardized data, and constructing a file and relation structure; modeling a file semantic space by adopting an improved TransH model, and generating semantic embedding representation; a semantic-time double-flow mechanism is introduced, and semantic flow and time flow joint representation is established; a multi-agent collaborative distillation system is constructed, and semantic, time and structure fusion is achieved; and calculating a comprehensive association score of the file and triggering incremental updating and synchronous correction. According to the method, the heterogeneous knowledge graph is constructed, and the improved TransH model, the semantic-time double-fluid mechanism and the multi-agent collaborative distillation system are combined, so that dynamic semantic modeling and accurate association analysis of multi-modal file information are realized.

The invention discloses a ship knowledge graph construction method based on a large model and a graph neural network, and the method comprises the steps: carrying out the format conversion and protection type partitioning processing of a professional document, and completing the recursive segmentation through a placeholder protection formula, a table and other structures according to the semantic hierarchy, and obtaining a text block suitable for the extraction of an entity and a relation; extracting domain entities by utilizing a large language model containing thinking chain cues, and positioning candidate entities in combination with an AC automaton so as to improve the coverage rate and accuracy of relation triple extraction; performing new entity backfilling and relation standardization on an extraction result, and constructing an initial knowledge graph with a consistent structure; and inputting the initial knowledge graph into a graph neural network model with directional expansion and a multi-scale decoder, and complementing a missing relationship through link prediction, so that node distribution and a relationship structure of the graph are more complete. According to the method, a continuous processing chain from text preprocessing, knowledge extraction to graph completion is formed.

The invention discloses an abnormal risk prediction method and device based on an assembly knowledge graph, equipment and a medium, and relates to the technical field of data processing. The method comprises the steps of obtaining multi-source heterogeneous data, performing association and alignment through a unified primary key set, and performing segmentation to generate a time slice data set; and constructing an assembly domain ontology and an assembly knowledge graph mode layer, instantiating the assembly domain ontology by using the time slice data set, and constructing an assembly knowledge graph snapshot sequence. And based on the assembly knowledge graph snapshot sequence, modeling and updating the influence relationship between variables in the assembly process to obtain an assembly process influence relationship structure. And in combination with the assembly knowledge graph snapshot sequence and the assembly process influence relation structure, coding is carried out, the state representation of the next time window is predicted based on the embedding representation of the historical time window, and an abnormal risk score is calculated. When the risk exceeds a threshold value, influence path tracking is carried out, controllable decision variables are screened, candidate intervention schemes are generated, the effects of the candidate intervention schemes are evaluated, and decision suggestions are output.

The invention discloses a chip IP interface information automatic analysis and connection modeling method, and relates to the technical field of integrated circuit verification, and the method comprises the steps: multi-source interface information collection: docking a design server through an FTP / SFTP protocol, automatically retrieving and capturing a target file, carrying out format preprocessing, and storing the collected original information through a database; heterogeneous information fusion analysis: extracting port physical attributes and generating machine executable rules by constructing a grammar analyzer and a semantic analyzer respectively, and automatically judging multi-source information contradictions by adopting a three-level judgment mechanism; structuring modeling of the connection relation: constructing a directed graph model of the IP nodes, the signal nodes and the constraint nodes, and automatically executing directed graph model verification; and through a standardized API interface docking verification tool, outputting a model file and a report file in standardized JSON-LD and DOT formats. Through cooperative work of cross-format acquisition, double-engine analysis, structured modeling and standardized output, full-automatic processing of IP interface information is realized.

The invention provides a power grid equipment fault prediction and maintenance scheme generation method. The method comprises the following steps: acquiring a data set; standardizing the data set to obtain a structured data set; based on the structured data set, defining a quintuple relation structure, and establishing a dynamic knowledge graph; constructing an encoder group; based on the data set and the dynamic knowledge graph, fusing the time sequence feature, the text feature and the graph feature by using an encoder group to obtain a fused feature; based on the fusion feature, obtaining a fault classification probability and a decomposition cost item; according to the fault classification probability and the decomposition cost item, determining a corresponding measure set and a feasible time window set in the dynamic knowledge graph to obtain a search space; and based on the search space, determining a target maintenance scheme so as to solve the problem that when an operation maintenance scheme is formulated by a power grid enterprise at present, systematic consideration on the equipment fault probability, the disposal cost and the resource configuration is often lacked, so that the scheme selection cost is high or execution conflicts exist.

The present disclosure relates to a data model establishing method and device, and a clinical auxiliary decision method and device. The data model establishing method comprises: field disassembling historical data according to different disease categories; structured processing is performed according to the fields obtained by disassembling to form a structured output rule; and a disease data model is established according to the structured output rule combined with clinical data. The present disclosure also relates to a data model establishing device, an electronic device and a computer readable medium, and a clinical auxiliary decision method, device, electronic device and computer readable medium based on the data model establishing method. The structured output rule is generated by structured processing of diseases, and the clinical data is saved in a relational structure in a database according to the rule to form a data model, so that the clinical data can be effectively utilized, the interconnection and intercommunication between data can be realized, and the research conversion rate can be improved.

The application relates to the field of natural language processing, and discloses a cross-language entity alignment method and system based on relationship semantic enhancement, which comprises the following steps: constructing entity initial embedding and relationship embedding representation of a cross-language knowledge graph; a key neighbor screening stage; screening key neighbor entities with structural matching and semantic matching based on an entity similarity matrix and a relationship similarity matrix and in combination with a preset semantic correlation threshold; neighbor-enhanced entity embedding updating; adopting a graph attention network to fuse semantic features of target entities and key neighbors thereof, and generating context-aware entity embedding; semantic consistency alignment propagation: first-order entity feature alignment and second-order relationship structure alignment are jointly optimized through a shared permutation matrix; dynamic seed expansion iteration: high-confidence unaligned entity pairs are dynamically identified according to Manhattan distances of entity embedding, entity pairs with distances lower than a threshold are added to a seed set, and the foregoing steps are iteratively executed until convergence.

The invention discloses a DAG-based block chain consensus method, which ensures the uniqueness and availability of blocks in a DAG by constructing a reliable broadcast algorithm under a new hybrid model. And meanwhile, throughput is improved by utilizing a DAG structure, a rapid confirmation path is fused and added to realize rapid confirmation and reduce delay, and then the activity of the protocol is ensured through a coin confirmation path. According to the method, a main-free design is introduced into a hybrid algorithm design, a multi-proposal architecture is adopted, block relationships of blocks of proposals are structured through a directed acyclic graph (DAG), and the activity and security of a system are ensured by using a random algorithm and a time hypothesis. In addition, by introducing a rapid confirmation condition, the delay is reduced while the throughput capacity of the protocol is ensured. The consensus of the blocks can be completed in the fastest four rounds of communication, and most of the blocks can complete the consensus of the blocks in the six rounds of communication, so that the low time delay of the block consensus process is ensured, and the responsiveness of the system is improved.

This invention discloses a training method, conversion method, apparatus, device, and medium for a language conversion model, belonging to the field of computer technology, and solves the problem of high training difficulty caused by the complexity of the model structure. The method includes: concatenating column name information from a natural language query sequence and a semantic matching association table into a long sequence; training a pre-constructed language conversion model based on the long sequence information to obtain a trained language conversion model. The language conversion model is used to determine the correspondence between the natural language sequence and the structured query language sequence, which includes selection clauses and conditional clauses. During the model training process, a first decoder is used to determine the prediction result of the selection clause based on its features, and a second decoder is used to obtain the prediction result of the conditional clause based on its features. This method can reduce the difficulty of model training.

This application belongs to the field of information retrieval technology and discloses a multimodal patent semantic retrieval method and system. By establishing relation anchor points, extracting local cross-modal relation representations around the anchor points, and verifying the cross-modal consistency of these relations, reliable cross-modal relations can be identified. Furthermore, technical features / components and successfully aligned cross-modal relations are constructed into a graph structure representation, and a graph matching algorithm is used to calculate similarity. As a result, the overall logic and details of the technical solution can be understood more effectively, and solutions containing only the same technical features but with different relational structures can be distinguished, thereby improving the accuracy and relevance of multimodal patent retrieval.

The invention discloses an integrated classifier multiplexing model fine tuning method and system, belongs to the technical field of artificial intelligence, and aims to solve the problems of feature structure damage and generalization ability reduction caused by hard masks after pruning of an existing model. The method comprises the following steps: constructing a virtual teacher model in a self-adaptive weighting manner based on the accuracy rates of local models of a plurality of clients on a public verification set; taking the pruned global model as a student model, performing multi-level knowledge distillation fine tuning by using a public data set, and covering output logic value alignment, middle layer feature map alignment and inter-class relation structure alignment; and repairing pruning damage through combined optimization of a composite loss function. The system comprises a data and model acquisition module, a virtual teacher model construction module, a composite loss calculation module and a parameter optimization module. According to the method, the performance and generalization ability of the pruning model can be effectively improved, and collaborative optimization of compression and precision is realized.

The invention discloses a data structure conversion method and device, electronic equipment and a storage medium, and relates to the technical field of databases. The method comprises the steps of obtaining a data table of a relation structure with an incidence relation; extracting at least one row data from the data table, and determining at least one graph node; obtaining foreign keys existing in primary keys in the data table, and determining graph nodes according to row data of the foreign keys and graph edges between the graph nodes generated by the fields with the incidence relation; aiming at each generated graph edge, generating a row positioning identifier according to the physical address of the row data of the graph edge index, and adding the row positioning identifier into the relation attribute of the graph edge; generating an adjacency list according to each graph node and each graph edge; and determining each graph node, each graph edge and the adjacency table as graph structure data of the data table. According to the embodiment of the invention, the graph structure data conversion accuracy and the depth query efficiency can be improved.

The invention discloses a graph data storage method and device, a graph data query method and device, equipment and a medium. According to the graph data storage method, after a graph data write-in request is received, the graph data write-in request is analyzed, and data of nodes or edges are extracted. And constructing a graph relation structure in the memory according to the extracted data of the nodes or the edges. And constructing at least one sub-graph according to the graph relation structure. And writing the at least one sub-graph into a disk in a log file form, and recording the unique identifier of the node or the edge of the at least one sub-graph and the file storage position of the at least one sub-graph in a metadata storage system. The graph data query method comprises the following steps: after receiving a query statement input by a user, analyzing the query statement to obtain a corresponding query parameter, after obtaining a unique identifier of a node or an edge according to the query parameter, querying in a disk and a sub-graph file of a metadata storage system according to a query plan, and outputting a query result. Therefore, the graph data storage and query efficiency can be effectively improved.

This invention discloses a method and system for updating entities in a knowledge graph. The method involves acquiring a new knowledge graph and an original knowledge graph; calculating a name attribute similarity matrix based on the name attribute of the new and original knowledge graphs; calculating an entity relation similarity matrix based on the entity relation structure triples of the new and original knowledge graphs; fusing the name attribute similarity matrix and the entity relation similarity matrix to obtain the entities corresponding to the new knowledge graph; and updating the original knowledge graph with the entities corresponding to the new knowledge graph. Advantages: Based on research on multi-attention entity alignment, this invention proposes a method for entity alignment in the fault diagnosis domain knowledge graph that combines the name attribute of long texts with relation structure similarity calculation. A knowledge graph update tool was developed based on this method. Through case testing and practical use, the accuracy of entity alignment and the efficiency of knowledge graph updates have been effectively improved.

An abnormal risk prediction method, device and equipment based on an assembly knowledge graph, and a medium are related to the technical field of data processing. The method comprises: acquiring multi-source heterogeneous data, correlating and aligning through a unified primary key set, and splitting to generate a time-sliced data set. An assembly field ontology and an assembly knowledge graph mode layer are constructed, the assembly field ontology is instantiated using the time-sliced data set, and a snapshot sequence of the assembly knowledge graph is constructed. Based on the snapshot sequence of the assembly knowledge graph, the influence relationship between variables in the assembly process is modeled and updated to obtain an assembly process influence relationship structure. In combination with the snapshot sequence of the assembly knowledge graph and the assembly process influence relationship structure, a state representation of a next time window is predicted by encoding and embedding representation based on a historical time window, and an abnormal risk score is calculated. When the risk exceeds a threshold, an influence path is tracked, controllable decision variables are screened, candidate intervention schemes are generated and their effects are evaluated, and a decision suggestion is output.

The invention discloses a federal traffic flow prediction method based on trajectory-driven directed relation modeling. The method comprises the following steps: 1) generating a basic hidden representation; (2) the directivity of traffic flow is enhanced; 4) mapping inflow and outflow representations to a shared potential space, and realizing directional semantic alignment by comparing loss; (5) constructing a directed relation graph subjected to minimum and maximum normalization; 6) generating time sequence features and uploading the time sequence features to a server; 7) dynamically updating the directed relation graph; 8) the server performs multilayer structure propagation on the time sequence features of the client by using the updated relational graph to obtain spatial enhancement features and returns the spatial enhancement features to the client; and 9) the server propagates client model parameters along the directed relational graph, acquires a personalized aggregation model and issues the personalized aggregation model to the client, so that the relational graph structure and the prediction model keep joint optimization in the federal training process. According to the method, a directed relation structure conforming to traffic direction dependence can be constructed, and fair, stable and accurate federal traffic flow prediction is realized.

The invention discloses a SysML model storage method fusing a key-value pair database and a graph database. The method comprises the following steps: S1, preprocessing and extracting model elements; s2, designing and storing a key value pair database; s3, designing and storing a graph database; s4, designing a graph-key value pair database association mechanism; and S5, verifying data integrity and boundary limitation. According to the method, the content information of the model elements is stored in the key value pair database in the JSON format, and the multi-dimensional semantic relationship among the model elements is structurally expressed in the graph database, so that a dual-mode database architecture with efficient read-write performance and rich semantic expression capability is constructed; a unified ID mapping mechanism and a semantic boundary verification rule are further introduced, decoupling storage and unified management of the model content and structure are achieved, and the technical blank of the SysML model in the aspects of extensible storage and relation modeling is filled up.

The invention relates to the technical field of agent application of a large language model, in particular to a large language model agent mass tool calling method and system based on knowledge graph retrieval enhancement, and the method comprises a tool graph construction stage, a tool semantic knowledge base construction stage and a tool retrieval calling stage. The method has the beneficial effect that the tool combination calling capability of the model is improved through the tool dependency relationship structure of the knowledge graph. Through vector retrieval and pre-retrieval of the knowledge graph, the number scale of tools for direct intention recognition of the large language model is reduced, and the error rate of multi-tool selection and tool parameter extraction slot positions of the large language model is reduced.

The invention discloses a time sequence knowledge graph inference model (HGR-GCN) based on a super relation and global historical perception, and aims to solve the problem that an existing TKG inference method is difficult to capture relation structural dependence and global historical information. According to the model, a super-relation graph is constructed for each timestamp historical sub-graph, adjacent entities and relation information are synchronously aggregated through a graph convolutional network, and a time evolution mode represented by a relation is learned in combination with a recurrent neural network module; and designing a global historical feature fusion mechanism, mining a complex combination rule of historical events, and generating comprehensive historical preference embedding as an additional feature channel for downstream prediction. Experimental verification on five public data sets including ICEWS18, YAGO and the like shows that the performance of the model in an entity prediction task is remarkably improved, the capturing capability of a time sequence knowledge graph on a complex historical evolution rule is effectively enhanced, and reliable support is provided for future event prediction.