247 results about "Semantic layer" patented technology

The invention discloses a small target detection method based on feature fusion and depth learning, which solves the problems of poor detection accuracy and real-time performance for small targets. The implementation scheme is as follows: extracting high-resolution feature map through deeper and better network model of ResNet 101; extracting Five successively reduced low resolution feature maps from the auxiliary convolution layer to expand the scale of feature maps. Obtaining The multi-scale feature map by the feature pyramid network. In the structure of feature pyramid network, adopting deconvolution to fuse the feature map information of high-level semantic layer and the feature map information of shallow layer; performing Target prediction using feature maps with different scales and fusion characteristics; adopting A non-maximum value to suppress the scores of multiple predicted borders and categories, so as to obtain the border position and category information of the final target. The invention has the advantages of ensuring high precision of small target detection under the requirement of ensuring real-time detection, can quickly and accurately detect small targets in images, and can be used for real-time detection of targets in aerial photographs of unmanned aerial vehicles.

A system and method for defining and enforcing a security policy. Security mechanism application specific information for each security mechanism is encapsulated as a key and exported to a semantic layer. Keys are combined to form key chains within the semantic layer. The key chains are in turn encapsulated as keys and passed to another semantic layer. A security policy is defined by forming key chains from keys and associating users with the key chains. The security policy is translated and exported to the security mechanisms. The security policy is then enforced via the security mechanisms.

A computer implemented method, apparatus, and computer usable program code for determining aggregate values of health data items from heterogeneously coded databases containing heterogeneously coded medical data. The data, in heterogeneous databases, is queried using a series of semantic layers including i) cascaded asymmetric association tables and ii) semantic search. The heterogeneously coded medical data items are translated into conformal dimensions and denominator files of combinations of disease data are derived. The denominator files of combinations of disease are aggregated based on a mapping of the coded medical and demographic conditions. The data is stored in a target data repository.

The invention discloses a smart voice interaction method and system. The method includes receiving voice data; performing voice recognition on the voice data and acquiring a voice recognition result;performing recognition rejection judgment on the voice recognition result according to a pre-constructed recognition rejection judgment model based on a semantic layer and acquiring a model output result; determining whether the voice data is man-machine interaction voice data or not according to the model output result; if the voice data is man-machine interaction voice data, performing semanticunderstanding on the voice recognition result and generating an interaction result according to a semantic understanding result, wherein the interaction result includes a response text. By utilizing the method and system provided by the invention, influence on man-machine interaction by noise-containing voice data can be reduced and error response of the man-machine interaction system is reduced.

The invention discloses a robot distributed type representation intelligent semantic map establishment method which comprises the steps of firstly, traversing an indoor environment by a robot, and respectively positioning the robot and an artificial landmark with a quick identification code by a visual positioning method based on an extended kalman filtering algorithm and a radio frequency identification system based on a boundary virtual label algorithm, and constructing a measuring layer; then optimizing coordinates of a sampling point by a least square method, classifying positioning results by an adaptive spectral clustering method, and constructing a topological layer; and finally, updating the semantic property of a map according to QR code semantic information quickly identified by a camera, and constructing a semantic layer. When a state of an object in the indoor environment is detected, due to the adoption of the artificial landmark with a QR code, the efficiency of semantic map establishing is greatly improved, and the establishing difficulty is reduced; meanwhile, with the adoption of a method combining the QR code and an RFID technology, the precision of robot positioning and the map establishing reliability are improved.

A system and method for defining and enforcing a security policy. Security mechanism application specific information for each security mechanism is encapsulated as a key and exported to a semantic layer. Keys are combined to form key chains within the semantic layer. The key chains are in turn encapsulated as keys and passed to another semantic layer. A security policy is defined by forming key chains from keys and associating users with the key chains. The security policy is translated and exported to the security mechanisms. The security policy is then enforced via the security mechanisms.

A method for operating a computer graphic system, the method comprising: inputting a media content object (MCO) into a feature extractor comprising semantic abstraction levels; extracting feature maps from said MCO on each of the semantic layers; selecting at least a portion of the MCO to be analysed; determining, based on the analysis of the feature maps from the portion of the MCO and the analysis of a previous state of a recognition unit, one or more feature maps selected from the feature maps of the semantic layers; determining a weight for each feature map; repeating the determining steps N times, each time processing, based on the analysis, each feature map by applying the corresponding weight; inputting said processed feature maps to the recognition unit; and analysing a number of said processed feature maps until a prediction about the portion of the MCO is output.

The invention relates to a semantic net based large scale offline data analysis framework. The large scale offline data analysis framework includes a data acquisition layer, a body layer, a data storage layer, a semantic layer, a data analysis layer and an application layer. A data source includes dynamic data and static data, and the static data includes data and database internal logic semantic and structure type. The static data is established into a body model in the analysis framework; the static data is extracted and modeled, and then the static data orients a user or an upper analysis task in a semantic service manner. The large scale offline data analysis framework can effectively improve the ability to organizing multi-source heterogeneous offline data and has a uniform interface to upper data; and application users or data analysis workers can access a lower data source through a semantic interface without knowing all the information of different data sources, and relevant data information is acquired. The large scale offline data analysis framework can effectively update the whole data source from a global perspective by correction of the body structure having changed content and update and inference service built in an application tool.

The invention discloses a semantic retrieval method oriented to field data. Semantic information of a certain field is described by using a body, so that a computer understands content inquired by a user more accurately, and retrieval on a semantic level is realized. A semantic retrieval model oriented to field data is provided in combination with the structural characteristics of a semantic network body on the basis of analyzing the prior art. The semantic relativity between a keyword and different examples is deduced by using the model according to the position of the keyword in the body, and a relativity weight is written into an example inverted list which corresponds to the keyword, so that the similarity computation complexity during semantic retrieval is lowered, and the retrieval efficiency is increased. According to the method disclosed by the invention, retrieval is raised from a grammatical level of simple keyword matching to a semantic level understandable for the computer, and the computer can understand the input keyword, so that retrieval intelligence is realized.

The invention discloses a Top-N movie recommendation method for performing weighted fusion on selected local models based on random anchor points. The method comprises the steps of obtaining eigenvectors of users and movies at the semantic level through an LDA topic model and a GBDT by utilizing movie text data; based on the eigenvectors, calculating Gaussian kernel similarity between the users and the movies; randomly selecting a plurality of (the users and the movies) anchor point pairs, and reconstructing a local training matrix for each anchor point pair in combination with the Gaussian kernel similarity between the users and the movies; performing training for each local training matrix by utilizing an SLIM as a basic recommendation model to obtain a corresponding local recommendationmodel; and finally, generating a final fusion recommendation model through weighted fusion among the local recommendation models. According to the recommendation method, the stability of the models is also maintained in a data sparseness scene, and the problem that a traditional single recommendation model is very easy to over-fit in the data sparseness scene can be effectively solved.

The invention provides a visual multi-database ETL integration method and system. The visual multi-database ETL integration method includes the following steps that source databases and target databases are connected; the SQL statements of source lists of the source databases are obtained through ETL matching of the source databases and the target databases; the SQL statements are optimized and executed, and the ETL data of the multiple source databases are obtained and introduced into target lists of the target databases. The visual multi-database ETL integration system comprises a database management system layer and a semantic layer, the database management system layer connects the source databases with the target databases, and the SQL statements of the source lists of the source databases are obtained through ETL matching of the source databases and the target databases; the semantic layer optimizes and executes the SQL statements, obtains the ETL data of the multiple source databases and introduces the ETL data to the target lists of the target databases. The visual multi-database ETL integration method and system reduce the complexity degree of integration of the multiple databases, improve the integration efficiency of the multiple databases, and reduce the risk of integration of the multiple databases.

The invention provides an improved grid topological semantic environment map construction method. A grid topological semantic hierarchy map can be established; a bottom layer grid map is used for a robot to perceive and understand the environment, an upper layer semantic map is used for a person to perceive and understand the environment, a middle topological map is used for establishing a corresponding relationship between the semantic name and the position coordinate, and the corresponding relationship is a connection bridge between the upper layer semantic map and the bottom layer grid map; according to the composition construction sequence, a grid layer, a topological layer and a semantic layer are sequentially constructed, according to the map searching sequence during the navigation process, the semantic layer, the topological layer and the grid layer are sequentially searched; and perception gap between grid data of the robot environment perception mode and the semantic conception of the people environment perception mode can be reduced.

The invention discloses a multi-source heterogeneous data fusion system which comprises a data source layer, a computing layer, a data layer and an analysis layer. The computing layer comprises a memory computing frame, a flow computing frame, a data warehouse, a data mining engine, a distributed type computing frame and a file system; the data layer comprises an SQL system, a NoSQL system and a caching system, and the analysis layer comprises a semantic layer and an OLAP engine. The invention further discloses a multi-source heterogeneous data fusion method which comprises the steps of S1, anairline company official website is transformed, and a user zipper table representing the user unique identity is obtained; S2, multi-source heterogeneous data are obtained, merged together and stored on a large data platform in a single user data manner; S3, application support is performed: the merged multi-source heterogeneous data are used for forming a user image; the formed user image is stored onto the large data platform in the NOSQL representation manner. According to the multi-source heterogeneous data fusion system and method, merging of the multi-source heterogeneous data is realized, and support is provided for scientific decision of the airline company.

The invention discloses a latent semantic image retrieval method of region-oriented multi-feature integration and multi-level feedback. It uses result list returned by the initial keyword search, extracting a variety of region-oriented images characteristics, constructing attribute-image matrix, using latent semantic indexing algorithm to get the semantic space of image sets and semantic features of each image, and then using similar images by users feedback to construct or update image query vector, searching again the semantic space, calculating image semantics features and images inquiries vector similarity, getting outcome sets by descending order, and repeatable retrieval. The invention takes full advantage of image content information, making up for the deficiencies of the keyword search, and through the region-oriented multi-feature integration, enhances image content information from the bottom physical layer to the object layer, then further enhances to the semantic layer by HCI feedback, thereby reducing the gap between the image bottom features and high-level semantic, and allowing Web image retrieval to get higher retrieval accuracy.

The invention aims at designing a method for generating an accounting voucher of an arbitrary business system document, which realizes business document semantic match of the accounting voucher to be generated by a peripheral arbitrary business system by the definition of a conversion semantic layer, and can realize directly generating the accounting voucher of a ledger accounting system of February 16 p, e2009 by the business document in the business system; in the conversion semantic layer two standard formats are provided and comprise: the match of a standard business data and a standard basic data as well as the match of a standard basic data and a subject assistant accounting; different voucher templates are defined by the business type of the standard business data, the business data of the peripheral arbitrary business system carries out a series of data match by the conversion semantic layer and generates the accounting voucher corresponding to the general ledger system by the voucher template defined by the conversion semantic layer.

Disclosed is a system and method of image-based space detection. The system includes an image selection module, a 3-layer detection mechanism and an optimization module. At least one image processing area that may affect space-status judgment is selected from plural image processing areas. The 3-layer detection mechanism having an observation layer, a labeling layer, and a semantic layer observes the information about the selected image processing area, associates with a local classification model, and adjacent local constraint model and a global semantics model to completely describe the probability distribution of the links among the three layers, and provide global label constraint information. The optimization module analyzes the probability distribution and global label constraint information, and generates an image-based optimized space detection result.

The invention relates to a method of adopting the multi-layer semantic annotation technology to analyze the intermediate result and detect the annotation information in layering mode based on the analyzed intermediate result during the malicious behavior detection. The model includes an annotation sub-module and a detection sub-module; wherein, the annotation sub-module includes a binary semantic layer, a linear instruction semantic layer, a control flow chart semantic layer, a process call chart semantic layer and an application level semantic layer. The annotation is processed towards the specified analysis target object in each layer. The detection sub-module detects the malicious behavior on the annotation information in layers based on the corresponding analysis strategy and the algorithm. The invention has the advantages of effectively decomposing the complexity of the malicious behavior detection, utilizing the advantages of different detection mechanisms, avoiding the inherent defects of different mechanisms, improving the adaptability and accuracy of the malicious behavior detection.

The invention relates to an automatic abstract generation method based on a concept semantic unit. The method comprises the steps that the concept semantic unit is utilized to serve as a carrier of semantic computation and document content representation; a semantic focus of document content is obtained through convergence of document semantic content; automatic generation of a document abstract formed by selecting abstract sentences representing the document content according to the semantic focus is achieved. According to the automatic abstract generation method based on the concept semantic unit, by means of a potential Latent Dirichlet Allocation (LDA) mode, modeling is conducted on a document theme, theme generation treatment is achieved, and the abstract sentences are selected according to the importance degree of the theme. In the computation process, a concept hierarchy network symbol is introduced, merging to semantic information is achieved by means of semantic hierarchy relations in the concept hierarchy network symbol, and the data sparsity problem caused by taking a word as a semantic computation unit is improved.

A business intelligence (BI) system which includes the ability to extend its functionality outside of the project life cycle by means of specific content. Complex multidimensional queries are interpreted as trees of atomic sub-expressions that are combined in a parse-tree-like structure to form the overall query. Each sub tree is valid in isolation when provided with the proper context. Any sub tree can be an expression template, stored as application content, which at generation time uses simple text substitution with instance specific parameters to produce multidimensional expression syntax. The system includes a sophisticated type system and semantic layer that hides the user from the complexities inherent in working with OLAP databases. A business intelligence expert can provide type and semantic cues for each expression template, held as content. The content expression templates are then exposed in the application primarily through a context menu that is filtered for appropriateness, but also in an explorer tree, toolbars, menus and submenus. The functionality from a users perspective is integral to the application. An iterative processing capability to complement these expressions is provided by means of OLAP database stored procedures held as application content. Building on the above, workflow content allows business users to extend the application by creating expert-system-like guided analyses and processes. Of key significance to this innovation is the concept that the expression templates, stored procedures and workflows are application content, and therefore redistributable and unshackled from the classic software development lifecycle and the cost and expertise associated.

The invention discloses a hierarchical semantic tree construction method and system for language understanding. The method mainly comprises the steps as follows: segmenting terms of a statement and loading a semantic knowledge base; recognizing all nodes of the statement according to an LV rule, and recognizing the level of the nodes according to semantic knowledge and term positions and collocations; generating a special node by punctuation at the end of the statement, and taking the special node as a root node of a semantic tree; merging the nodes according to generated node information, recognizing semantic side chunks of the statement, and taking a level-0 semantic side as a child node to be hung on the root node; circularly traversing all child nodes of the statement till no low-level semantic side exists, and taking the child nodes as leaf nodes to be hung on the child node. According to the hierarchical semantic tree construction method and system, under the condition that no syntactic resource exists, the semantic structure tree is obtained through semantic information and the term positions and collocations only, so that a computer can enter a deep semantic layer of a natural language, various processing of the natural language can be finished on the basis of understanding, the first step of semantic understanding of the natural language is realized, and the hierarchical semantic tree construction method and system can be applied to information retrieval, automatic abstraction, machine translation, text categorization, information filtration and the like.

A semantic infrastructure (SI) system is provided to interface with an information asset management tool, such as a document management system. The SI is a semantic layer that links together structured, unstructured and tacit information. The SI determines a set of valid terms associated with remaining facets based on the term selection for a first facet and in accordance with predefined facet relationships. The present invention links the content of documents to an enterprise view and to a corporate file plan. The SI includes a database storing assets each associated with a plurality of facets. Embodiments of the present invention provide a unified approach, i.e., an approach that can provide file classification, taxonomies, metadata, thesauri and data model functionalities, to better describe and organize information assets using a single database employing a faceted classification system. A stored metadata instance is mapped to an externally supported metadata format.

The invention relates to a field device information management system based on semantics and OPC UA and belongs to the combined field of the semantic network and the industrial Internet of things. Theintegral configuration of the system comprises a sensing layer, a network layer, a semantic layer and an application layer, wherein the sensing layer comprises a bottom field device and an OPC UA server, the network layer comprises each network system accessed by the OPC UA, the semantic layer comprises an OPC UA device information acquisition end, a device information management domain knowledgeontology model, a device information semantic annotation module, a semantic reasoning and query module, an ontology database and a semantic rule file, and the application layer comprises a device running state management module, a device operator management module, a device machine account management module, a device spare part management module, a device maintenance management module and a device repair management module. The management system is advantaged in that semantic integration of the heterogeneous device information is realized, through semantic reasoning and query, the relatively rich semantic knowledge in the field can be acquired, and needs of each functional module of the system are satisfied.

The invention relates to a video interaction event analysis method and device base on sequence space-time cube characteristics. The method includes the steps of dividing a monitoring video into a plurality of space-time cube sequences on the basis of a detection tracking result of the monitoring video, extracting object tracks, the appearance and local movement descriptors in each space-time cube, forming characteristic segments through the extracted descriptors, reconstructing the characteristic segments in all the space-time cubes to establish the sequence space-time cube characteristics, and conducting interaction event classification detection through the sequence space-time cube characteristics. The device comprises a preprocessing module, a video sequence dividing module, a space-time cube characteristic extraction module, a space-time characteristic reconstruction module and a sequence characteristic classification module. According to the method and the device, high-level semantic layer description is achieved for content of the monitoring video, length-variable sequence characteristic classification is achieved through a multi-core support vector machine based on a dynamic time alignment kernel function, and therefore intelligent detection is achieved for monitoring video stream interaction events.

An analytic visualization, such as a chart or graph, embedded into a container. The container is embedded into a portal, such as a web page, that is viewable by a viewer device. To update the analytic visualization, an update server receives a data request from the viewer device or from a container host server that hosts the container. The update server generates a data processing instruction based on the data request, which it sends to data sources. The data sources store a full dataset, and are configured to extract a processed dataset from the full dataset based on the data processing instructions. The update server receives the processed dataset from the data sources, applies one or more semantic layer operations to the processed dataset, and generates a visualization update based on the result. The update server then transmits the visualization update to the viewer device or container host server.

Embodiments of the invention are generally directed to a system and method for enterprise data objects. An enterprise data object is a container of data that can be used in a wide range of technical and semantic layers within a business system. In particular, enterprise data objects provide high-level interfaces that can be used to access and manipulate complex and structured data.

The invention discloses a semantic recognition method and system based on a knowledge graph. The method comprises the following steps: constructing the knowledge graph in advance, wherein the knowledge graph comprises a phonetic layer, a word layer, a presentation layer, a semantic layer and an intention layer; receiving input information; converting the input information into phonetic units; determining a word unit related to each phonetic unit and a presentation unit related to each word unit; determining a semantic unit related to each presentation unit; selecting a selected semantic unit according to each semantic unit and a relation between a forerunner group located in front of a corresponding position of the corresponding presentation unit in the input information and a following group behind the corresponding position of the corresponding presentation unit in the input information; determining an intention unit related to each selected semantic unit, and selecting the selected intention units from the intention units according to a relation between each intention unit and the corresponding selected semantic unit; determining that a selected intention set composed of the selected intention units is an intention corresponding to the input information. Therefore, the semantic recognition method and system can be used for carrying out semantic recognition on all natural languages.

A computer readable storage medium includes executable instructions to define a semantic domain describing a data source. At least one base dimension is established as a data model object for the semantic domain. At least one base measure is specified as a data model object for the semantic domain. At least one of a calculated measure or a calculated dimension is described as a data model object for the semantic domain.

The invention provides a method for organizing and applying design knowledge supporting product innovation, comprising the following steps of (1) dividing the design knowledge supporting the product innovation into principle knowledge, comprehensive knowledge and field knowledge; the comprehensive knowledge and the field knowledge are collectively called design fact knowledge; (2) expressing the design fact knowledge using a physical characteristic layer, expressing the principle knowledge by using an innovation property layer, and expressing the incidence relation between the physical characteristic layer and the innovation property layer using a knowledge semantic layer; (3) constructing an internal semantic network side to organize the principle knowledge; (4) constructing an external knowledge resource network side to organize the design fact knowledge; (5) constructing the relation between the internal semantic network side and the external knowledge resource network side; and (6) constructing a design knowledge base application system facing the product innovation, and providing brand new, tracking and improved three different knowledge application models for product innovation design according to the requirements of design tasks.