137 results about "Link model" patented technology

A computer-implemented pattern recognition method, system and program product, the method comprising in one embodiment: creating electronically a linkage between a plurality of models within a classifier module within a pattern recognition system such that any one of said plurality of models may be selected as an active model in a recognition process; creating electronically a null hypothesis between at least one model of said plurality of linked models and at least a second model among said plurality of linked models; accumulating electronically evidence to accept or reject said null hypothesis until sufficient evidence is accumulated to reject said null hypothesis in favor of one of said plurality of linked models or until a stopping criterion is met; and transmitting at least a portion of the electronically accumulated evidence or a summary thereof to accept or reject said null hypothesis to a pattern classifier module.

A computer-implemented pattern recognition method, system and program product, the method comprising in one embodiment: creating electronically a linkage between a plurality of models within a classifier module within a pattern recognition system such that any one of said plurality of models may be selected as an active model in a recognition process; creating electronically a null hypothesis between at least one model of said plurality of linked models and at least a second model among said plurality of linked models; accumulating electronically evidence to accept or reject said null hypothesis until sufficient evidence is accumulated to reject said null hypothesis in favor of one of said plurality of linked models or until a stopping criterion is met; and transmitting at least a portion of the electronically accumulated evidence or a summary thereof to accept or reject said null hypothesis to a pattern classifier module.

The invention discloses a combined scheduling method of satellite-ground measurement and control resources of a low-mid-orbit satellite constellation based on orbit determination constraint satisfaction, and belongs to the field of combined scheduling of the satellite-ground measurement and control resources. The method comprises the steps of (1) building a visible geometry model, (2) building a visible link model, (3) building the constraint condition of orbit determination, (4) carrying out preprocessing, (5) determining a resource distribution principle, (6) building a combined scheduling model of the satellite-ground measurement and control resources, and (7) carrying out algorithm solution. Compared with the prior art, the method has the advantages that the scheduling result is reasonable, the scheduling process is efficient, and the scheduling method is high in universality. The air-and-ground-based integrated scheduling of the measurement and control resources is achieved. The transition from ground-based measurement and control to air-and-ground-based combined measurement and control in China is achieved.

The invention discloses a system for controlling the production process of a petrochemical enterprise on the basis of manufacturing execution system (MES) workflow. The system comprises an MES production flow model base, an MES production flow information database, a workflow server, a production control terminal workstation and a data interaction interface, wherein the MES production flow model base is used for storing MES production flow link models and an MES whole production flow model for describing topological structures among the MES production flow link models; the MES production flow information database is used for storing production execution data in each link of the MES production flow; and the workflow server is used for invoking the MES production flow model base and the MES production flow information database to execute an MES workflow task. The invention also discloses a control method of the system for controlling the production process of the petrochemical enterprise on the basis of MES workflow. The system and the method provide activity execution information of panoramic production links for operators in a graphical mode so as to control the progress of the production flow in real time and improve the production efficiency.

The invention discloses an emulation system for a PaaS (Platform-as-a-service) cloud platform. A static model comprises a node model, a topology model, a link model, an application model and a user model, wherein the node model is used for determining computing resources, communication resources and application deployment conditions; the topology module is used for determining role definition of each node and topology connecting conditions; the link model is used for determining communication link attribute of the node; the application model is used for determining computing resource spending, communication resource spending and time delay for application logic and application processing; the user model is used for determining user requirement distribution and user requirement achieving conditions; the dynamic model comprises a control model, a protocol model and an event model, wherein the control model is used for determining operation logic and layout relationship of application and node as well as selecting a mode of processing nodes for business request after reaching the business request; the protocol model is used for determining an interacting framework of the PaaS cloud platform; and the event model is used for determining dynamic events with uncertainty generated in a PaaS cloud platform operating process. The invention discloses an emulation method based on the system. According to the emulation system for the PaaS cloud platform and the emulation method for the PaaS cloud platform disclosed by the invention, emulation load conditions of the node in the PaaS cloud platform can be determined.

The invention relates to a method and a system for recognizing the human face. The method comprises the following steps: a human face which is captured in an image is detected; the profile of the human face is extracted from the detected human face, and a plurality of human face mark feature vectors are obtained; a dynamic-link attribute graph is generated according to the obtained human face mark feature vectors; and the comparison of the generated attribute graph and the attribute graph in a human face database is carried out to determine whether the two are matched or not. The system comprises a server subsystem and one or more client subsystems. The system is a full-automatic and intelligent human face detection and verification system. The method adopts a dynamic-link model of the human face profile graph to realize the efficient storage and recognition of the human face graph, so as to achieve the high correct recognition rate.

A second internet protocol network is logically connected to a packet data network connection provided between a user equipment and a first internet protocol network over a radio access network, the second internet protocol network located on a data path from the first internet protocol network to the user equipment. The first internet protocol network represents the highest level internet protocol point of attachment to the packet data network connection. Router advertisements are sent from the second internet protocol network to the user equipment over the radio access network via the packet data network connection.

A unifying network model with a structure and architecture configured to address security, interoperability, mobility, and resource management, including priority and quality of services is provided. The network of the network model is structured as a hierarchical mesh network, with dynamically generated routing tables. The configuration of the network model optimizes routing and distributes communication load. Every device on the network is capable of being both an endpoint and a forwarder of communications. The network model may include underlying networks that are represented with one of two models, the link model or the star model. The nodes are organized in a hierarchical relationship structure to optimizes throughput. The model may include a cryptographic method of dynamically assigning local network addresses.

A learning apparatus for learning time series data in a link model including a plurality of input time series pattern storage networks and a plurality of output time series pattern storage networks with nodes of the input time series pattern storage networks linked to nodes of the output time series pattern storage networks, includes a learning unit for updating in a self-organizing manner each of the plurality of input time series pattern storage networks and updating in a self-organizing manner each of the plurality of output time series pattern storage networks and a link relationship updating unit for updating a link relationship between each node of the output time series pattern storage network and an input winner node, and updating a link relationship between each node of the input time series pattern storage network and an output winner node.

A differential game anti-interception maneuver penetration / precise striking guide method includes the following three steps of firstly, carrying out modeling of a three-body differential game model, wherein a warring three-party dynamics link model, a warring three-party linearized warring model and a three-body differential game model are included; secondly, carrying out dimension reduction processing on an original three-body differential game model and obtaining the new three-body differential game model; and thirdly, solving the new three-body differential game model based on the optimal control theory and obtaining the optimal control law of a guided missile, wherein solving of the Hamiltonian function of the new three-body differential game model and the optimal control problem corresponding to the new three-body differential game model is included. By means of the three steps, the establishing process of the three-body differential game model, the conversion process of the new model and the old model and the solving process of the differential game model are described, and finally, the differential game anti-interception maneuver penetration / precise striking guide method is obtained.

The invention discloses a dynamic financial knowledge graph construction method based on reinforcement learning and transfer learning. The method comprises the following steps of: 1) constructing a financial knowledge graph for structured data and semi-structured data of each selected listed company, inserting the entity reference corresponding to the financial entity in the graph into a financialentity database, 2) obtaining a financial entity data set for the unstructured data related to the selected listed company, 3) training a financial entity identification model by using the financialentity data set and the standard entity identification data set, 4) generating a financial entity link data set, and training a financial entity link model by using the financial entity link data set,5) finding a financial entity corresponding to each entity reference in the unstructured data in the financial knowledge graph by using the trained financial entity link model and updating the financial knowledge graph, and 6) performing financial entity relationship extraction from the unstructured data by using a financial relationship extraction model and updating the financial knowledge graph.

Aimed at the current situation of poor on-orbit absolute radiometric calibration precision, the invention provides an on-orbit absolute radiometric calibration method based on a benchmark satellite. Sunlight with appropriate intensity is reflected onto a to-be-calibrated remote sensor by a diffuse reflection plate on a benchmark satellite, the influence of atmospheric conditions on a radiation transfer path is got rid of, an accurate apparent radiance is obtained, and thereby an accurate absolute calibration coefficient is obtained. In the invention, the sun, the benchmark satellite and the to-be-calibrated remote sensor are integrated into a geocentric coordinate system, so that the distance between each two of the three and the incidence angle and emergence angle of sunlight on the surface of the diffuse reflection plate on the benchmark satellite can be conveniently obtained, a radiation transfer link model from the sun to the benchmark satellite and to the to-be-calibrated remote sensor is created, and finally, calibration is complete.

Displacements of respective joints corresponding to respective joint elements (J9 and the like) of a rigid link model (S1) representing a two-legged walking mobile body (1) are sequentially grasped. Also at the same time, values, in a body coordinate system (BC), of an acceleration vector of the origin of the body coordinate system (BC) fixed to a waist (6) as a rigid element, a floor reaction force vector acting on each leg (2), and a position vector of the point of application of the floor reaction force vector are sequentially grasped. With the use of the grasped values, joint moments respectively generated in an ankle joint (13), a knee joint (14), and a hip joint (9) of each leg (2) are sequentially estimated based on an inverse dynamics model using the body coordinate system. The estimation accuracy of the joint moments of the leg can be enhanced by reducing arithmetic processing using tilt information of the two-legged walking mobile body relative to the gravity direction as much as possible.

An apparatus comprising a plurality of flow model circuits, a categorizer circuit, and a plurality of link model circuits. Each of the flow model circuits may be configured to generate a respective flow measurement signal in response to data packets received from a plurality of links on a network. The categorizer circuit may be configured to present a plurality of link measurement signals, each in response to one or more of the plurality of flow measurement signals. Each of the link model circuits may be configured to generate a respective congestion estimation signal in response to (i) a respective one of a plurality of model parameter signals, (ii) a respective one of the plurality of link measurement signals, and (iii) a respective one of a plurality of score signals. Each of the congestion estimation signals may represent a congestion level of a respective one of the links.

The invention discloses an active above-knee prosthesis terminal sliding mode control method based on a disturbance observer. The method includes the steps that basic information data of a patient is collected in an off-line mode, and a data report is generated; on the premise of not changing the basic structure and the interaction mode, the lower limbs of the human body are simplified into a two-link model to be analyzed, the lower limb kinetic model of the supporting period and the swing period is obtained aiming at the difference between the swing period and supporting period motion form in the human body walking process; the disturbance observer and a terminal sliding mode controller are designed for the lower limb kinetic model. The sliding mode variable structure control theory is introduced into control over active above-knee prosthesis, while advantages of disturbance resisting capacity and quick response of traditional sliding mode control are reserved, the buffeting degree of controlled quantity is weakened, and the tracking errors of a model system can be converged to zero within limited time. The method has the advantages of being high in control accuracy, high in disturbance resisting capacity and good in timeliness.

A method and apparatus for obtaining a set of optimized angles of arrival for a corresponding set of radio links. The set of radio links model a radio environment of a wireless unit operating at a particular location within in a radio system. Each radio link represents a different propagation path between the wireless unit and transmitting antenna operating within the radio system. Each optimized angle of arrival represents an angle of arrival of one radio link with reference to the wireless unit. Each probe antenna of a set of probe antennas is positioned at a corresponding angle of the set of optimized angles of arrival. A corresponding set of probe radio signals is transmitted from the set of probe antennas.

Computer method, apparatus and system chains model-to-model transformations. In a series of model transformations, there are respective bridges before and after each model transformation. For each pair of consecutive model transformations in the series, the respective bridge (i) receives a model output from a first model transformation of the pair, the received model being in the respective output model format of the first model transformation of the pair, (ii) prepares the received model as input to a second model transformation of the pair, including preparing the received model to be in the respective input model format of the second model transformation of the pair, and (iii) inputs the prepared model to the second model transformation of the pair. The series of model transformations and respective bridges provide chaining of the model-to-model transformations. Each model transformation in the series is able to be separately configured.

The embodiment of the invention provides an entity linking method, an entity linking device, electronic equipment and a medium, belongs to the technical field of multimedia, and aims to improve the performance of entity linking. According to the embodiments of the invention, the method comprises the steps of: receiving a to-be-identified text input by a user, determining a to-be-identified entityword in the to-be-identified text, and searching candidate entry information associated with the to-be-identified entity word from a preset knowledge base; inputting the to-be-identified text, the to-be-identified entity word and the candidate entry information into an entity linking model; encoding the to-be-identified text and the to-be-identified entity word by an encoding part of the entity linking model to obtain the feature vector of the to-be-identified text, encoding each piece of candidate entry information associated with the to-be-identified entity word to obtain the feature vectorof each piece of candidate entry information, calculating the association degree between the feature vector of the to-be-identified text and the feature vector of each piece of candidate entry information, outputting the association degree, and then determining the candidate entry information with the highest association degree as entry information linked to the to-be-identified entity word.

The technology disclosed relates to representing users in different contexts within customer relationship management (CRM) environments. In particular, it relates to creating a person object model for each user in the CRM environment by linking a plurality of user records in the CRM environment that is separately created by different organizations and departments. The person object model also incorporates online social identities of the users, according to one implementation. This inclusive linked model provides a comprehensive view of the users to handlers of the CRM environment, thus facilitating an enhanced overall CRM experience.

Displacements of respective joints corresponding to respective joint elements (J9 and the like) of a rigid link model (S1) representing a two-legged walking mobile body (1) are sequentially grasped. Also at the same time, values, in a body coordinate system (BC), of an acceleration vector of the origin of the body coordinate system (BC) fixed to a waist (6) as a rigid element, a floor reaction force vector acting on each leg (2), and a position vector of the point of application of the floor reaction force vector are sequentially grasped. With the use of the grasped values, joint moments respectively generated in an ankle joint (13), a knee joint (14), and a hip joint (9) of each leg (2) are sequentially estimated based on an inverse dynamics model using the body coordinate system. The estimation accuracy of the joint moments of the leg can be enhanced by reducing arithmetic processing using tilt information of the two-legged walking mobile body relative to the gravity direction as much as possible.

The invention discloses an efficient routing method for an orthogonal circular orbit configuration satellite constellation. According to the invention, the method comprises the steps: proposing a principle of proximity, a principle of approach and an equatorial orbit satellite active periodic updating mechanism according to a constellation link model to ensure the rationality and the cheapness ofa network link; proposing a link priority algorithm based on hop count prediction on the basis of the link model; enabling a satellite to send out a data packet according to the priority of each linkthrough sequential decisions of a link actual on-off state, a link pseudo-off state caused by a non-backhaul strategy, a destination node direction, a link congestion degree and a link propagation delay factor; carrying out discussion in four situations according to the categories of the source satellite node and the destination satellite node, and providing a hop count prediction method for sending data packets from different links for each situation according to satellite numbers and state information.

The invention provides a simulation method and system for a large-scale scene of an electric power communication optical transmission network. The simulation method comprises the following steps: A: establishing a link model with transmission attributes, and simulating the optical transmission network by using the link model; B: simulating a fault; C: obtaining fault information, calculating a newservice forwarding path after network self-healing protection according to the fault information and pre-stored forwarding path information, and calculating corresponding new transmission attributesof the link model according to a new service forwarding path; and D: updating the link model according to the new transmission attributes. According to the simulation method and system for the large-scale scene of the electric power communication optical transmission network provided by the invention, the service forwarding path in the electric power communication optical transmission network is abstracted as the link model with small fluctuation of delay and error code, the self-healing protection of the fault in the electric power communication optical transmission network is abstracted as the switching of the link model, so that the simulation of the large-scale scene can be completed in a single simulation software.

The invention provides an entity linking method for a Chinese knowledge graph question-answering system. The method comprises the following steps: firstly, performing joint embedding on words and entities in a training corpus to obtain joint embedding vectors of the words and the entities; for an input text of the Chinese knowledge graph question-answering system, firstly, recognizing entity reference items in the input text, and determining a candidate entity list according to the entity reference items; and constructing an entity link model based on an LSTM network, performing vector splicing on the entity representation vector and the entity reference item representation vector to obtain a similarity score of the entity reference item and the candidate entity, and finally obtaining a score rank of the candidate entity, thereby selecting the candidate entity with the highest score as a target entity corresponding to the entity reference item. According to the method, the defect of link model training data redundancy caused by diversity of user questioning modes is effectively solved, and words with similar semantics can be replaced and used in the context, so that the link effectiveness is improved, and the accuracy of a question and answer system is improved.

The invention discloses a method for checking and optimizing a single point fault of an electric power communication service. The method comprises the steps of establishing a physical link model of acommunication network; analyzing the total amount of service information, and establishing the service set to be checked; obtaining the service logical channel routing information according to the service information to be sorted out, obtaining the service physical channel routing information, and searching for a single point of hidden trouble; reconstructing the logical link model of communication network, adding optical path resource occupancy ratio information, marking the physical link co-routing situation in the logical link, and establishing the logical link co-routing set; and routing optimization of a single point of failure is carried out. The method of the invention has the advantages that the hidden trouble of the single node is quickly identified and found, and the optimizationsuggestion of the optimal route is provided according to the requirements of different users through the optimal route calculation module, so as to accurately eliminate the hidden trouble of the single point of the fault; the optimization proposal can provide reference to the staff, avoid missing the situation, save a lot of manpower and material resources and improve work efficiency, and is highin degree of automation.

A computer-based real-time streaming system under packet erasures wherein created messages can be decoded within a fixed delay form their creation is presented. Various code construction methods and corresponding hardware implementation for different cases of erasure link models are also presented.

Computer method, apparatus and system chains model-to-model transformations. In a series of model transformations, there are respective bridges before and after each model transformation. For each pair of consecutive model transformations in the series, the respective bridge (i) receives a model output from a first model transformation of the pair, the received model being in the respective output model format of the first model transformation of the pair, (ii) prepares the received model as input to a second model transformation of the pair, including preparing the received model to be in the respective input model format of the second model transformation of the pair, and (iii) inputs the prepared model to the second model transformation of the pair. The series of model transformations and respective bridges provide chaining of the model-to-model transformations. Each model transformation in the series is able to be separately configured.

The invention provides a satellite intelligent data transmission method based on remote sensing state estimation and a system thereof, and the method comprises the steps: estimating the satellite-ground data transmission link loss in real time according to uplink power detection and link model estimation; calculating the maximum data transmission capability of data transmission according to the satellite-ground data transmission link loss; identifying an interested remote sensing region through preprocessing target detection and hot spot region designation; adjusting a remote sensing data compression ratio and compression quality according to the obtained maximum data transmission capability and the data volume of the interested remote sensing area; caching the preprocessed target detection data, the original remote sensing data and the compressed remote sensing data, and selecting and storing data information needing to be downloaded according to the maximum data transmission capability and the interested remote sensing area data volume; reading to-be-downloaded data, and adjusting a satellite data transmission link according to the to-be-downloaded data volume and the transmission link loss. The code rate, the coding mode, the modulation mode and the transmitting power of the remote sensing data link are optimized, and the remote sensing data transmission capacity is improved.

The invention discloses a simulation method oriented to LDoS (Low-rate Denial of Service) and LDDoS (Low-rate Distributed Denial of Service), and mainly aims to solve the problem of high simulation computation overhead in the current research of LDoS and LDDoS based on a discrete event network simulation technology. The method comprises the following implementation steps: (1) reading a network simulation topology; (2) reading UDP (User Datagram Protocol) attack flows; (3) reading normal TCP (Transmission Control Protocol) flows; (4) constructing a network simulation topology scene based on a simplified queue and a link model; (5) reducing the pulse amplitude of the UDP attack flows to 1 / P, and configuring an attack flow simulation application; (6) reducing the quantity of the TCP flows to 1 / P, and configuring a TCP flow simulation application; (7) running a simulation scene based on a discrete event network simulation technology, and recording the simulation result of the LDoS / LDDoS; and (8) amplifying the recorded simulation result of the LDoS / LDDoS by P times to form a final simulation output result. By adopting the simulation method, the simulation computation overhead can be reduced greatly on the premise of ensuring the LDoS / LDDoS simulation authenticity, and simulation deduction of a large-scale LDoS / LDDoS attack-attack-defense strategy can be supported.

The invention discloses a data preprocessing method based on a block chain intelligent contract and machine learning. Data preprocessing on nodes is completed through cooperation of a block chain technology and a machine learning technology. The method specifically comprises the steps of collecting Internet of Things data, and standardizing the Internet of Things data after storing the Internet ofThings data to local nodes; C; creating a machine learning algorithm file on the local node, wherein the machine learning algorithm file is used for preprocessing the Internet of Things data; S; storing the data into a block chain; T; training a learning algorithm by using the linked data to obtain a model coefficient, and storing the model coefficient in a local node; Firstly, a machine learningalgorithm expression is written into an intelligent contract, and then an obtained model coefficient is stored into a block chain by calling the intelligent contract; P; processing the to-be-processed data by using the linked model; A; and storing the processed data into a block chain. According to the method, a machine learning algorithm is added into an intelligent contract of a block chain, and the data mining capability of the block chain is improved by utilizing machine learning.

The invention discloses a passive damping control method and system of a single-phase LCL grid-connected inverter. The method comprises steps as follows: an established mathematical equation is converted into an open-loop transfer function of the single-phase LCL grid-connected inverter, the open-loop transfer function of the single-phase LCL grid-connected inverter is decomposed into a form of aproduct of an LCL filter model, a delay link model and a current regulator model, each link Bode image is drawn, the system stability is analyzed according to the open-loop transfer function of the control system and each link Bode image, system current regulator parameters are designed after constraint conditions calculated according to a nyquist diagram are added to the design of LCL filter parameters, and passive damping control of the single-phase LCL grid-connected inverter is realized. The system stability can be improved under the influence of delay of a digital system, damping controlis not required, additional sensors are not required, and the hardware cost of the control system is reduced.