110 results about "Link model" patented technology

The invention belongs to the field of artificial intelligence and the field of financial science and technology, and relates to a knowledge graph construction method and device based on artificial intelligence, computer equipment and a storage medium. Performing preprocessing to obtain an insurance knowledge text; performing entity extraction on the insurance knowledge text based on an entity extraction model to obtain an insurance entity; performing attribute extraction on the insurance knowledge text and the insurance entity based on an entity attribute extraction model to obtain an attribute value of the insurance entity; performing relation extraction on the insurance knowledge text based on a relation extraction model to obtain an insurance relation; based on the entity link model, performing data mapping processing on the ontology model by using the insurance entity, the attribute value of the insurance entity and the insurance relationship to obtain a mapping result; and performing knowledge storage processing on the mapping result based on the graph database to obtain a target knowledge graph. In addition, the target knowledge graph may be stored in a blockchain. According to the invention, the construction efficiency and the construction accuracy of the knowledge graph are improved.

The invention relates to the technical field of wireless communication, in particular to a reachable rate optimization method based on an intelligent reflecting surface, which comprises the following steps of: firstly, constructing a channel model in an application scene, namely designing a system reachable rate optimization model on the basis of establishing a base station-user, base station-IRS and IRS-user three-link model; then, a phase shift matrix of the IRS and an emission covariance matrix of a base station serve as joint optimization variables, a population is divided into an elite subpopulation, a suboptimal population and a common population through a subpopulation grouping strategy of an improved particle swarm algorithm, each particle represents the composed variable, a dynamic cooperation strategy is proposed to calculate the speed in each subpopulation, and the speed in each subpopulation is calculated; and updating individual optimization and group optimization, and continuously carrying out iteration to search a reachable rate optimal solution of the user. According to the method, local search can be finely carried out while large-range global search can be carried out, the calculation complexity is remarkably reduced, and the IRS transmission effect is improved.

The invention provides a modeling and KC analysis method for a double-front-axle balanced suspension. The modeling method comprises the steps that the structural form of the double-front-axle balanced suspension is determined; constructing a plate spring model of the double-front-axle balanced suspension; under the structural form, hard point coordinates in the double-front-axle balanced suspension are determined; and establishing the hard point coordinates in the plate spring model to obtain a corresponding balance connecting rod model, and assembling the balance connecting rod model to obtain an assembly body model of the double-front-axle balanced suspension. The KC analysis method comprises the following steps: newly adding a first-bridge and second-bridge reverse jumping working condition in a KC analysis project; and performing KC analysis on the assembly model of the double-front-axle balanced suspension on the basis of the first and second axle reverse jumping working condition to obtain a KC performance index of the assembly model under the first and second axle reverse jumping working condition. By means of the method and device, the defects existing in modeling and KC analysis modes of double-front-axle balanced suspensions and vehicle suspensions in the prior art are overcome.

The invention provides an electric power system evolution deduction method based on hierarchical collaborative simulation, and belongs to the field of novel electric power system long-term planning. The method comprises the following steps: constructing an electric power system evolution boundary; simulating an electric power system distant view design; embedding electric power system source network load storage collaborative planning simulation of short-term operation simulation to form a current-stage power grid planning scheme; recording an evolution stage scheme; according to the method, the evolution process is summarized into simulation of three levels of long-term, middle-term and short-term according to the characteristics of different links of the power system, multi-factor and multi-time-scale coupling simulation is realized through modular modeling and solving of each layer of simulation, and the modeling complexity and accuracy are considered, so that the method can be used for large-scale system evolution deduction; and meanwhile, the modular modeling and solving support the self-defined assembly of a main link model and a solving method of the evolution system so as to keep better openness, and the expandability and adaptability of the evolution system are remarkably improved.

The invention discloses a multi-modal content collaborative recommendation method based on a large model. The method comprises the steps that input is constructed, text information adopts a title of a to-be-recommended article, visual information obtains a feature vector of an article picture through a visual coding module, and collaborative information obtains hidden space vector representation of a user and the article through a collaborative coding module; information coding: integrating the recommendation data according to a template of the prompt information, coding the recommendation data and inputting the recommendation data into an LLM (Logical Link Model) to generate recommendation; and multi-stage fine tuning: enabling the model to master the input multi-modal information through three-stage fine tuning in combination with a task adapter. According to the method, differentiated interests and tendencies embodied when a user faces different modal information are emphasized, multi-modal information can be integrated, text information, visual information and collaborative information are uniformly incorporated into a large model, a task adapter module is introduced for the visual information and the collaborative information, and the user experience is improved. The modal characteristics highly related to the specific recommendation task can be extracted, so that the recommendation performance is improved.

The invention relates to the technical field of satellite communication, in particular to a VLEO-LEO satellite network quantum key distribution method based on MILP. The method comprises the following steps: S1, establishing a satellite-ground link model containing atmospheric attenuation characteristics of a VLEO satellite and an LEO satellite, and analyzing a coupling relationship between a key generation rate and link stability; s2, constructing an MILP-ARA algorithm based on time slot division, realizing optimal path selection under multiple service priorities in the discretization topology by introducing a flow conservation constraint, a key capacity constraint and an end-to-end time delay constraint, and dynamically balancing the link quality and the key utilization rate by using an adaptive weight factor; and S3, large-scale constellation simulation verification is passed. According to the VLEO-LEO satellite network quantum key distribution method based on MILP provided by the invention, the heterogeneous constellation cooperative relay model is constructed, and the static topology of time slot division and the dynamic key management mechanism are combined, so that core challenges such as inter-satellite link high dynamics, key resource limitation and link attenuation are solved.

The invention discloses a photoelectric imaging system multi-component optimization method based on full-link model driving. The method comprises the following steps: determining a parameter value range of to-be-optimized sensor parameters of multiple components in a target photoelectric imaging system; the to-be-optimized sensor parameters comprise focal length, aperture, pixel size and integration time; based on a pre-constructed full-link simulation model of a target photoelectric imaging system and a target genetic algorithm, performing iterative optimization on the parameter value range of the to-be-optimized sensor parameters to obtain target parameter values of the to-be-optimized sensor parameters under aberration-free and stray-free conditions; and determining a target aberration range and a target stray irradiance range of the target photoelectronic imaging system based on the target parameter values of the optimized sensor parameters. According to the method, the synergistic effect and the mutual coupling restriction relation between different component parameters can be integrated in the optimization process, and the performance of the optimized target photoelectric imaging system is improved.

The invention provides an active networking control modeling method for a high-proportion renewable energy system, which belongs to the field of energy, and comprises the following steps: constructing an active power output variable quantity model of a converter under small disturbance so as to establish a linear relation between the active power output of the converter and a load power variable quantity and a linear relation between the active power output of the converter and a phase angle difference variable quantity of the converter; constructing a VSG control model, and establishing a mathematical model of a converter control link; vSG represents a virtual synchronous generator; constructing an active power-frequency response model for parallel operation of a plurality of VSGs; and a prediction and correction link model is constructed to accurately reflect the actual dynamic behavior of the system. The method provides theoretical support for active networking control performance optimization.

The application discloses a kind of space multiplexing time slot allocation method and system based on underwater acoustic channel conflict matrix, first according to underwater acoustic transmission link model, the link signal receiving success probability in underwater acoustic network is calculated;Second, according to underwater network topology information and link signal receiving success probability, generate channel conflict matrix;Then, construct time slot allocation matrix is used to mark node occupies time slot situation, from time slot conflict situation, time slot occupancy and allocation uniformity angle, to time slot allocation construct multi-objective optimization function;Finally, according to underwater acoustic channel conflict matrix and time slot allocation matrix, through global search, find the optimal solution of time slot allocation.The application comprehensively considers the allocation of underwater channel resources from three aspects of time slot conflict, time slot utilization and time slot allocation uniformity, establishes an effective fitness evaluation mechanism, can make full use of channel resources, and improves underwater acoustic communication performance.

The present invention discloses a method for humanoid robot standing with variable load based on model reference adaptive control, comprising the following steps: constructing a three-link model of the humanoid robot, deriving its dynamic equation, and obtaining the joint torque expression of the three-link model; using the calculated joint torque as the feedforward torque, and forming the joint torque control quantity together with the output of the PD controller, and inputting it into the control system of the humanoid robot; calculating the update rate of the dynamic parameter estimation value, updating the feedforward torque, until the control error of the joint angle and the joint angular velocity is eliminated, and completing the construction of the model reference adaptive control algorithm when the humanoid robot stands with both legs side by side; and debugging the humanoid robot standing with variable load. The present invention uses feedforward control, and calculates the joint torque required to support the gravity load through the dynamic equation of the three-link model when the legs stand side by side, and directly adds it to the output end of the PD controller, which not only shortens the control time, but also ensures the joint control accuracy.

The present invention provides a method, system, readable storage medium and computer equipment for scheduling unmanned aerial vehicles. The method includes dividing a flight area to obtain a plurality of spherical areas; designing link stability indicators of a plurality of unmanned aerial vehicles, obtaining link information transmission index values ​​between a plurality of unmanned aerial vehicles; optimizing a channel model; selecting corresponding links based on link information transmission index values ​​and the optimized channel model, obtaining a hypergraph link model between a plurality of unmanned aerial vehicles based on a disjoint multipath model, and constructing a conflict graph through the intersection relationship between a plurality of unmanned aerial vehicles; obtaining the weight of each flight path according to a plurality of paths in the hypergraph link model and the conflict graph, and scheduling a plurality of unmanned aerial vehicles based on an optimization algorithm and the weight of each flight path. The present invention can reduce the energy consumption of unmanned aerial vehicles, avoid conflicts in information transmission, and solve the problems of scheduling of unmanned aerial vehicle networks and scheduling mismatch of unmanned aerial vehicle network changes.

The invention provides a transformer substation secondary system loop abnormity analysis positioning method and system, and belongs to the field of power system automation. The method comprises the following steps: establishing a network topology structure model and a machine account model; the two models are perfected through a system configuration description file, a logic connection link model, a physical connection link model and a corresponding relation are generated, and a secondary circuit digital model is constructed; receiving link breakage alarm information of the secondary loop model, determining a logic connection link associated with the abnormal signal according to the logic connection link model, mapping the logic connection link to a physical connection link model, traversing all possible abnormal physical ports in a communication link, and counting the occurrence frequency of alarms; all physical connection link models associated with the abnormal physical port of the communication link are searched, the involved physical port is analyzed, and the most probable fault node is determined by calculating the abnormal weight. According to the invention, automatic analysis of the abnormity of the secondary system loop of the transformer substation is realized, and the decomposition result accuracy and the positioning accuracy are improved.

The invention particularly relates to a cascade communication resource graphical interaction method based on an intelligent terminal. According to the cascade communication resource graphical interaction method based on the intelligent terminal, geographic position information of a whole park is rendered and presented in a GIS map page; selecting a park plane section object of the park, and rendering and presenting relative position information of cables, optical cables, tube wells and access machine room resources in the park plane graph; establishing a link model according to the optical cable and cable information, and displaying link routing, connection information and a filling rate; and recording the total length, the number of used cores and the number of residual cores of the optical cable and the electric cable in the park, and carrying out data statistics and analysis. According to the cascade communication resource graphical interaction method based on the intelligent terminal, the efficiency and accuracy of communication resource management are improved, so that the communication resource management is more intelligent, automatic and efficient, and new vitality is injected to sustainable development of a park.

The invention provides a method, device and equipment for modifying BIM link model data and a storage medium, and relates to the technical field of BIM design, and the method is applied to a BIM link scene and comprises the steps that an axis net of a BIM link model is loaded in a current document, when the axis net is in a model state, execution of editing operation on the axis net is forbidden, and in a current view, editing operation on the axis net is forbidden. And in response to a first state switching operation for the axis net in the current view, switching the axis net from a model state to a view state, and when the axis net is in the view state, in response to an editing operation for the axis net, modifying the axis net view data corresponding to the current view in the second data area without changing the axis net model data in the first data area. The invention provides the method. In a BIM model link scene, on the premise of not changing basic data of an axis network in a link document, flexible display and controlled editing of the axis network of the link model in a specific view of the current document are realized.

[Problem] To provide an information processing system and the like capable of optimizing an entire manufacturing line. [Solution] According to one aspect of the present invention, an information processing system is provided, the information processing system comprising a processor configured to execute following steps by reading a program, wherein the processor executes: a model acquisition step to acquire a pre-process model in which a pre-process for generating an intermediate product from a raw material is simulated, and a post-process model in which a post-process for generating a final product from the intermediate product is simulated; a condition acquisition step to acquire a process condition for each of the pre-process model and the post-process model; and an inference step to acquire, as an inference value, at least one of a characteristic value of the intermediate product, a characteristic value of the final product, a state value of a device used in the pre-process, and a state value of a device used in the post-process, by inputting the process conditions into a linked model in which at least the pre-process model and the post-process model are linked.

The present invention discloses a structured data modeling method, modeling system, and management method. The method comprises: constructing a knowledge processing link model based on structured data; constructing a knowledge label matrix model based on the knowledge processing link model; performing instantiation determination on the knowledge label matrix; and arranging the structured data and the knowledge label matrix after instantiation determination to perform data set entity placement, thereby completing structured data modeling. The method effectively handles the processing link from data to information and then to knowledge, and automatically completes the construction of the knowledge label system and the knowledge link during the modeling process, laying a solid data model foundation for the subsequent research and development of applications such as knowledge search and intelligent recommendation.

The invention relates to the technical field of artificial intelligence, in particular to a graph neural network and dynamic relationship interactive graph entity linking method, which comprises the following steps of: constructing a training data set for completing a multi-modal knowledge graph entity linking task; extracting multi-modal features mentioned from the entities in the training data set to obtain input features; a multi-modal entity link model based on multi-scale dynamic relation interaction is constructed, and model parameters are initialized; inputting the input features into a network to obtain multi-scale output features; optimizing and updating model parameters according to the loss function; repeating the steps of inputting the input features into the network, obtaining the multi-scale output features and optimizing and updating the model parameters according to the loss function until a specified round is reached; and inputting test data into the trained model, selecting the data with the maximum similarity as a final link result according to a link strategy, and by fully utilizing the features of different scales, avoiding the problem of single-scale feature loss and enhancing the multi-modal entity link effect.

The invention discloses a special protocol message cross-space semantic association reasoning method and system for an electric power monitoring system, and aims to solve the problems that protocol analysis and system security events are isolated and advanced threats are difficult to associate in existing monitoring. The method comprises the following steps: jointly analyzing a protocol transmission time sequence through Fourier transform, wavelet transform and an intelligent time sequence prediction model, and constructing a fine-grained time characteristic baseline; a sensing agent is deployed on a host, a database, security and protection equipment and other layers, and formatted security events are collected; and designing an intelligent association engine fusing graph representation learning and rule reasoning, carrying out deep semantic association on an abnormal protocol mode and a cross-level security event, and realizing accurate threat alarm based on a business behavior full-link model of reinforcement learning self-optimization. According to the method, intelligent cross-space reasoning from network protocol microscopic abnormity to system macroscopic threat is realized, and the precision, the intelligent level and the adaptive capability of network security situation awareness of the intelligent substation are greatly improved.

The embodiment of the invention relates to the technical field of knowledge maps, and provides an entity linking method, a related device and a storage medium, the method comprises the following steps: obtaining an initial search text input by a user, the initial search text comprising at least one target word; calling an entity link model to obtain at least one candidate pair matched with the initial search text, wherein the candidate pair comprises a standard word associated with at least one target word on at least one of semantic features and font features; and setting a candidate entity corresponding to the standard word in the candidate pair as a target entity corresponding to the target word in the initial search text according to a corresponding relationship between the standard word and the entity. According to the scheme, reasoning can be carried out by utilizing knowledge of similar instances in a training set during entity prediction, so that the prediction result of the model is corrected, and a correct entity is easier to output.

The invention provides a large model operation method, system and device and a storage medium, and the method comprises the steps that a model service generates model registration information according to model information input by a user and the architecture of a server where the model service is located, and sends the model registration information to a container engine; the container engine obtains a model file corresponding to the model registration information from a file warehouse according to the model registration information, generates a mirror image file corresponding to the model registration information based on the model file, and sends the mirror image file to the model service; the model service triggers the target container to run according to the mirror image file, and when the target container runs, a running link of the target container is generated; and the model service obtains the operation link, and triggers the target large model executed by the operation link to operate according to an instruction of a user. According to the method, automatic deployment and operation of the target large model can be realized, and the workload of manual installation and configuration is reduced.

The embodiment of the invention provides a task system testing method and device and a storage medium. When the task system is tested, a link model of a target task system is determined, and a link node to be tested is determined from the link model. According to the mapping relation between the link node and the test case, obtaining the test case corresponding to the link node to be tested from a test case library; according to the task processing logic of the to-be-tested link node, arranging a test case corresponding to the to-be-tested link node to obtain a test case combination; and executing the test cases in the test case combination to test the link node to be tested. In this way, the task module to be tested in the task system can be visually determined based on the link model of the task system. According to the mapping relation between the link node and the test case, the test case can be determined and executed, so that automatic testing of a complex task system comprising a plurality of task modules is realized, and the test cost is reduced.

The invention discloses a crane obstacle avoidance function anomaly detection method and system based on time sequence behavior recognition, and relates to the technical field of anomaly detection, and the method comprises the following steps: collecting multi-source state data of a crane in an operation process, and constructing an obstacle avoidance time sequence behavior link model; generating an expected execution event set of the obstacle avoidance action based on the obstacle avoidance time sequence behavior link model; extracting an actual execution event sequence corresponding to the expected execution event set from multi-source state data; and performing time sequence offset trend comparison on the expected execution event set and the actual execution event sequence, and judging whether the event is an obstacle avoidance function abnormal event or not. According to the invention, the time sequence behavior link model between the obstacle avoidance triggering condition and the action response is constructed, and the offset comparison is carried out on the expected execution event and the actual execution event; the problems that in the prior art, obstacle avoidance function abnormity cannot be accurately recognized based on the time dependency relation of the whole trigger-execution process, and the obstacle avoidance performance degradation trend cannot be judged are solved.

The invention belongs to the technical field of artificial intelligence, and relates to an LLM-based business order automatic input method and related equipment thereof, and the method comprises the steps: recognizing a customer business intention through business communication data; if the intention is the expected intention, binding a business entry related service component for the bill entry agent; generating a business input planning scheme; and according to the business entry planning scheme, business sheet entry is automatically carried out. When the method is applied to a financial service communication or negotiation scene, for example, when an insurance communication salesman or a customer service performs service communication with a customer, the big language processing model is utilized to identify the service intention of the customer, when the service intention is an expected intention, the big language processing model is adopted to perform service sheet input planning, and finally, the business sheet input planning is performed. According to the input planning scheme, business order input is automatically carried out, and compared with manual input, the input time can be saved, the business order input efficiency can be improved, manual input errors caused by manual input can be avoided, and the business order input accuracy can be improved.

A finite-time anti-interference control method for a universal pneumatic flexible manipulator includes the following steps: The universal pneumatic flexible manipulator is simplified to a link model using the D-H method, and kinematic analysis is performed to obtain the kinematic analysis model and the position of the end point of the universal pneumatic flexible manipulator; based on the kinematic analysis model of the universal pneumatic flexible manipulator, a dynamic model of the universal pneumatic flexible manipulator is established using the Euler-Lagrange method; the uncertainty of the universal pneumatic flexible manipulator is considered as a disturbance, and a second-order mathematical model of the pneumatic flexible manipulator is established; a finite-time extended state observer is designed to estimate the disturbance, and based on the disturbance estimate, a finite-time backstepping controller is designed to compensate for the influence of the disturbance on the system; to address the model uncertainty, a finite-time control method is designed to compensate for the influence on the position control accuracy of the pneumatic flexible manipulator, exhibiting strong robustness and high control accuracy.

The present invention proposes a method for power system evolution deduction based on hierarchical collaborative simulation, which belongs to the field of new power system long-term planning. The method includes the following steps: constructing the power system evolution boundary, simulating the power system vision design, and embedding the power system source-grid-load-storage collaborative planning simulation with short-term operation simulation to form the current stage power grid planning scheme and record the evolution stage scheme and make evolution termination judgment to connect each stage in series to form a long-term evolution result. The present invention summarizes the evolution process into three levels of simulation: long-term, medium-term and short-term according to the characteristics of different links of the power system. Through modular modeling and solution of each layer of simulation, multi-factor and multi-time scale coupling simulation is realized, and the modeling complexity and accuracy are taken into account, so it can be used for large-scale system evolution deduction; at the same time, modular modeling and solution support the customized assembly of the main link models and solution methods of the evolution system while maintaining good openness, significantly improving the scalability and adaptability of the evolution system.

The invention discloses a bulk commodity transportation and sales full-link big data processing method, and particularly relates to the technical field of big data processing. The method comprises the following steps: collecting real-time data of production, transportation, storage and sales nodes, and constructing a heterogeneous data source set; constructing a weight mapping matrix based on the data frequency and the credibility; fusing and constructing a space-time unified full-link model; extracting a dynamic feature sequence of the transportation batch, wherein the dynamic feature sequence comprises transportation time consumption, a path fluctuation deviation coefficient and a congestion abnormal coefficient; generating an event response vector through anomaly detection and fluctuation recognition; a causal backtracking feedforward prediction model is constructed in combination with historical sales data, and a sales prediction result is dynamically adjusted; according to the optimized transportation path, generating a scheduling suggestion set; incremental updating is carried out on the full-link model to realize closed-loop optimization; the method can effectively improve the prediction precision, scheduling efficiency and link stability of bulk commodity transportation and sales.

The embodiments of this specification provide a product data processing method, device and equipment, which can be used in the financial field or other fields. The method includes determining the target task based on the user's selection of attributes in the task attribute page; obtaining task information corresponding to the target task; when it is determined that the target product is a hook model according to the task information, obtaining indicator data corresponding to the hook model; wherein the hook model is used to calculate the profit and loss of the product in different stress scenarios; wherein the indicator data corresponding to different stress scenarios are different; based on the hook model and the indicator data, determining the result offset of the target product in the specified stress scenario. The embodiments of this specification can solve the problems of incomplete stress testing projects and inflexible measurement between products, and ensure the normal operation of the bank under various pressures.

Embodiments of the present application provide a text-based entity linking and recognition method, an electronic device, and a storage medium. The method includes: determining text data as training data; inputting the training data into an entity linking model; processing the training data by a translation module of the entity linking model to obtain corresponding output features; splicing the output features with prior knowledge features to obtain fused features; and making a judgment on entity linking for the fused features to determine the linked entity as the output result of the entity linking model. By integrating prior knowledge in a knowledge graph into the model and comprehensively predicting entity linking, the accuracy of the entity linking model can be improved.

The application belongs to the technical field of cross-language entity linking, and discloses a cross-language entity linking method, system, medium, equipment and terminal.The cross-language entity linking method comprises the following steps: searching for candidate entities and constructing an entity name index; constructing an entity linking model based on different clues; performing vector embedding and fusion based on attribute-based entity linking; and performing multi-clue entity linking based on co-occurring entities.The application uses three kinds of clues to realize entity linking, the clues are entity attributes, co-occurring entities and context descriptions, and neural network structures such as LSTM, CNN and GCN are combined to perform word embedding representation and network modeling.Compared with traditional entity linking technology, the application has stronger semantic representation ability and the linking method has cross-language capability.Through analyzing different forms of information, the application realizes a multi-clue cross-language entity linking algorithm with three different structures, and completes accurate linking of news text cross-language entities such as persons and institutions and a given knowledge base.

The invention provides a manipulator control system based on visual identification and touch sensing, and relates to the technical field of manipulator control. Comprising the steps that a manipulator control module obtains joint parameters, constructs a multi-finger connecting rod model and obtains a control parameter instruction; and the visual identification module analyzes the underwater target image by adopting an SSD300 algorithm to obtain the position and shape features of the object. The path optimization module controls grabbing according to the control parameter instruction, adjusts the instruction according to the pressure distribution data to obtain a feedback control signal, and optimizes the path according to the object characteristics. And the control module operates the manipulator, and the underwater target object is grabbed according to the feedback control signal and the optimized path. According to the method, through the generated control parameter instruction, the control precision of the manipulator can be improved, the manipulator can reach the target position more accurately and conduct grabbing operation, meanwhile, the SSD300 algorithm is used for introducing underwater target detection, and underwater target detection and positioning can be achieved with high precision.