27 results about "Dependent model" patented technology

The invention discloses a reinforcement learning training method and system for relieving illusion of a multi-modal large model, and belongs to the field of reinforcement learning training of a multi-modal large language model. Firstly, a planning and visual description generation step is introduced in an early stage to guide a model to perform structured reasoning, then a grouping relative strategy optimization algorithm is used, reward values are calculated for multiple candidate responses generated by the model after cold start, and particularly, a visual perception reward mechanism is set. The reward mechanism evaluates the consistency of the generated text description and the visual information by using an external large language model. Then, based on a vision description attention score advantage distribution method, learning of the model on key vision signals is dynamically enhanced, and the perception ability of the model on the vision signals is improved; and finally, the perception and reasoning performance of the model is further improved by adopting multiple rounds of rejection sampling and supervised fine tuning. The scheme does not depend on a model architecture, the extra overhead is small, the illusion problem caused by early image-text inconsistency is effectively solved, and the accuracy and the reliability are improved.

The invention discloses a multi-granularity data cascade updating method and device based on a dynamic topological graph and a medium, and relates to the technical field of data processing. The method maintains a multi-level data dependency model (DAG) in memory. The system recursively updates a target node by monitoring the change of a source node and utilizing an incremental propagation algorithm. For a high-concurrency scene needing isolation verification, the method introduces an overlay shadow topology mechanism: on the premise of not destroying a main graph structure and not copying a total graph, only instantiating a shadow copy of an affected node based on a Context ID, and constructing an overlay Mapping pointing to a father node; when the propagation path is calculated, the numerical value of the shadow node is preferentially read based on the context, and the propagation path of the original node is logically blocked in the current context. The invention further provides a path convergence mechanism and rendering frame synchronization technology based on topology in-degree, and the problems of resource competition, data consistency and front-end rendering flicker under the complex dependence network are effectively solved.

The invention relates to the technical field of software engineering automation of digital data processing, and discloses a game development automation method based on a multi-agent system. The method is used for solving the problem that a machine-readable cross-product dependence model and consistency verification are lacked in a traditional method. According to the method, based on a task list freezing project context, a constraint packet and a threshold aperture, and based on a demand account book and registration demands, constraints, codes, configuration, resources and other entry abstracts, version information and strong and weak reference relationships; the multiple agents only submit change packets in a temporary storage area by writing tokens and occupying fragmentation write-in gating, and an influence range report binding verification list is generated for each change packet; boundary compliance, reference integrity, compilation verification, determinacy operation verification and acceptance assertion verification are sequentially carried out on the temporary storage area snapshots, and when conflicts occur, rollback or confinement repair is triggered to output traceable and acceptance deliverables based on reproducible evidence judgment.

The invention discloses a learner cognitive diagnosis method oriented to an intelligent education scene, and belongs to the technical field of cognitive diagnosis and education data analysis. According to the method, concept-level multi-dimensional modeling is carried out on the ability and exercise difficulty of a learner by constructing multi-dimensional representation of concept perception so as to describe mastering characteristics of the learner on different cognitive levels; and meanwhile, distinguishing a pre-correction dependency relationship and a semantic approximation relationship, establishing a relationship-perceived concept dependency model, and deducing a potential exercise-concept association structure according to the relationship-perceived concept dependency model. Further, the annotated exercise-concept incidence matrix and the inference incidence matrix are fused in a unified diagnosis layer, and comprehensive evaluation of the knowledge mastering state of the learner is achieved. According to the method, an end-to-end mode is adopted for optimization, the fineness, knowledge coverage and stability of cognitive diagnosis can be effectively improved, better prediction performance and generalization ability are shown on multiple real education data sets, and the method is suitable for intelligent education application scenes such as learning analysis and personalized teaching.

The application discloses a cross-media model error attribution and auxiliary correction method based on causal decoupling, which comprises the following steps: image segmentation is performed on an input image, and the input image is divided into a plurality of sub-regions; a target function is solved according to the sub-regions, the minimum key region which supports the generation of a target word element to the greatest extent and removes the generated ability is sorted according to the contribution to the generation decision, and an attribution saliency map is generated; an influence score is calculated according to the change of the generation probability of the target word element in the process of inserting the sub-regions in the ordered subset; the ordered subset, the attribution saliency map and the influence score are evaluated from the dimensions of fidelity, positioning ability and error correction guiding ability, and an evaluation result is obtained; the evaluation result is obtained without relying on the internal gradient, attention weight or activation map of the model, the dependence on the internal structure of the model is reduced, the regional attribution of an arbitrary word element set is realized, and the relative dependence of the generation process on visual evidence and language prior is quantified.

A computer-implemented method for ascertaining an item of environment information is provided. The item of environment information relates to material in surroundings of a third article and / or an interaction of the third article with the material. The method includes receiving an X-ray image and an item of sensor information, and determining model parameters or limiting possible parameter values of the model parameters of a three-dimensional model of the third article as a function of the X-ray image in order to specify an X-ray-dependent model. The three-dimensional model describes a three-dimensional shape and / or pose of the third article as a function of the model parameters. The method includes ascertaining the item of environment information as a function of the X-ray-dependent model, where the item of environment information and / or the X-ray-dependent model additionally depends on the item of sensor information. The item of environment information is provided.

The invention relates to the technical field of reliability management of electromechanical equipment, in particular to an electromechanical equipment reliability evaluation method based on a switching Markov process and fractional Brownian motion, which comprises the following steps: constructing a switching Markov process model, the description module is used for describing a multi-stage degradation process experienced by equipment in a life cycle and state transition thereof; based on the external working condition parameters and the switching Markov process, constructing a stage dependency model; establishing a nonlinear degradation model of the electromechanical equipment based on the fractional Brownian motion process, and introducing a random effect into the model to represent individual differences; and constructing a two-stage parameter estimation method, and sequentially estimating unknown parameters in the switching Markov process and the degradation model. The performance degradation of the electromechanical equipment is accurately modeled, and the nonlinearity, randomness, individual difference, long-term memorability and stage dependence of the multi-stage degradation of the electromechanical equipment are effectively captured, so that the reliability of the electromechanical equipment is accurately evaluated.

This invention relates to the field of data processing technology for predictive purposes, and discloses a method for adjusting the power generation performance of an energy system throughout its entire life cycle. The method determines the parameter boundaries of a wind-solar-storage power supply system, divides the entire life cycle planning period into multiple consecutive life stages, and determines the load demand input for the entire life cycle. It sets a baseline power generation performance at the initial commissioning stage, calculates the degradation maintenance level for each stage, establishes a monotonic mapping relationship between the degradation maintenance level and maintenance intensity, constructs a stage-dependent model for calculating available power generation capacity coupled with degradation and maintenance, and obtains available power supply capacity parameters for each stage of the entire life cycle. It generates a maintenance time series distribution through Monte Carlo simulation and constructs a set of life cycle power generation performance evolution scenarios. With the goal of minimizing the total cost throughout the entire life cycle, a mixed-integer linear programming model is constructed to obtain the optimal capacity configuration and adjust the power generation performance throughout the entire life cycle. This method solves the problems of ignoring degradation maintenance and inaccurate planning, achieving the goal of accurate and reliable life cycle data at a low cost.

The invention provides a super junction structure parameter optimization method and apparatus, a terminal device and a storage medium. The method comprises the steps of establishing a temperature dependence model of each physical parameter in a super junction structure; wherein the temperature dependence model is used for describing the change rule of the physical parameters along with the temperature; establishing constraint conditions according to a planar junction theory and an avalanche breakdown principle; wherein the constraint conditions comprise a collision ionization integral breakdown condition and a buffer layer boundary condition; obtaining target information based on the design target of the super junction structure; wherein the target information comprises breakdown voltage, depth-to-width ratio and temperature; solving a performance parameter optimization value of the super junction structure under the target information according to the temperature dependence model and the constraint condition; wherein the performance parameters comprise depletion layer thickness, doping concentration and specific on-resistance. According to the method, the optimization calculation of the performance parameters of the super junction structure can be realized in a larger temperature range.

The application discloses a kind of based on double-layer dependent perception Web3D application mobile edge cache method, first according to the data of content object in Web3D application mobile network, double-layer dependent relationship modeling is carried out, the logical dependence model and space dependence model between Web3D application content object are obtained, then based on double-layer dependent model, mobile edge cache problem with the minimum system expected loading total delay as target is constructed, and edge server carries out caching and replacement to the resource module in each content object according to the cache decision variable matrix obtained by solving.The logical dependence model and space dependence model between Web3D application content object are obtained by double-layer dependent relationship modeling in the application, and caching decision is carried out based on the above two models, and the hit rate and resource loading efficiency of Web3D application mobile edge cache are improved.

The invention provides a cloud configuration item intelligent optimization system and method, and relates to the technical field of cloud configuration data optimization. Comprising the following steps: constructing a dynamic configuration relation graph by mapping configuration item data to different hierarchies of a rule dependence model, generating a multi-modal evidence body and a dynamic trust chain rule based on the dynamic configuration relation graph and key attributes of configuration items, constructing a causal dependence graph through instantiation, calculating a robustness index for the causal dependence graph, and establishing a dynamic trust chain rule. Performing conflict resolution arbitration according to the state of the arbitrated configuration item; the data accuracy and the self-healing capability are improved through automatic conflict resolution, the real-time performance and the consistency of configuration optimization are enhanced through causal dependence analysis, and the optimization efficiency is improved while the reliability is ensured.

The application relates to the technical field of computer vision and deep learning, and particularly discloses a fabric defect detection method based on a visual saliency model, which utilizes a convolutional neural network to construct a spatial detail extraction branch, accurately captures the edge and texture details of local tiny defects through a self-adaptive weighted context coordination mechanism, simultaneously introduces a visual state space model to construct a global information extraction branch, and efficiently establishes a long-distance dependence model of a global background with the aid of a two-dimensional space selective scanning technology. On this basis, a bidirectional feature fusion decoder is used to perform deep cascading and dynamic integration on local details and global contexts, thereby effectively suppressing the interference of regular texture backgrounds, enhancing the feature saliency of irregular defects, and realizing high-precision and high-efficiency automatic detection of fabric defects in a low-contrast scene.

The application discloses a Go language library migration method based on semantic dependency graph and complexity scheduling, and comprises the following steps: S1, source code is acquired, and a unified semantic dependency graph is generated; S2, a minimum migration unit is generated; S3, multi-dimensional complexity quantification and risk modeling calculation are carried out on each minimum migration unit, a risk score is generated, all minimum migration units are sorted according to the risk score from low to high, and a migration priority queue is constructed; S4, according to the migration priority queue, the migration unit with lower risk is preferentially executed, and a migration state is recorded; S5, a test verification process is carried out; and S6, migration scheduling is carried out. The application realizes automatic migration planning based on program analysis, graph theory structure and complexity scheduling algorithm, does not depend on model training or external knowledge enhancement mechanism, and has high engineering applicability.

A computer-implemented method for determining environmental information (22) based on at least one two-dimensional X-ray image (23) depicting at least a section of a third object (24-26) located within a patient (42), and at least one sensor information (27) based on at least one measurement value from at least one sensor (28-31) of the third object (24-26), wherein the environmental information (22) relates to material (32-34) in the vicinity of the third object (24-26) and / or an interaction of the third object (24-26) with this material (32-34), comprising the steps: - Receiving the X-ray image (23) and the sensor information (27), - Determining model parameters (35) or restricting the possible parameter values ​​of the model parameters (35) of a three-dimensional model (36) of the third object (24-26) depending on the X-ray image (23) to specify an X-ray-dependent model (37), wherein the three-dimensional model (36) describes a three-dimensional shape and / or pose of the third object (24-26) depending on the model parameters (35), - Determining the environmental information (22) as a function of the X-ray image-dependent model (37), wherein on the one hand the environmental information (22) and / or on the other hand the X-ray image-dependent model (37) additionally depend on the sensor information (27), and - Providing the environment information (22).

The application discloses a data-driven nonlinear multi-agent system online cooperative control method, relates to the technical field of multi-agent system cooperative control, considers an affine nonlinear multi-agent system with external disturbance and unknown system model under a non-directional connected graph, and realizes consistent terminal value bounded control of the system. The scheme comprises the following steps: considering a discrete time affine nonlinear multi-agent system with external disturbance and unknown system model, a distributed state feedback controller is designed; based on an online updating data collection mechanism, state data and input data of the system are collected, and a data-based system parameterized representation is constructed; a state-dependent model is constructed, a data-based distributed robust controller is solved online, and consistent terminal value bounded control of the unknown model affine nonlinear multi-agent system is realized.

The invention provides a three-dimensional variation data assimilation method based on a flow dependence model, and the method comprises the following steps: S1, carrying out the preprocessing of multi-source data and the setting of a region-element adaptive time window: carrying out the preprocessing of multi-source meteorological data, and dynamically setting a time window [tau] based on the characteristics of meteorological elements and the characteristics of a region; s2, constructing and training a flow dependency model: constructing a deep learning flow dependency model of a CNN-LSTM hybrid architecture; and S3, constructing a cost function to carry out gradient solution. According to the three-dimensional variation data assimilation method based on the flow dependence model, a deep learning flow dependence model is introduced, a meteorological element association model in a small-scale time dimension is accurately constructed through a high-precision multi-feature extraction capability, traditional static prior information is replaced, and the new three-dimensional variation method has dynamic time adaptability; time windows are dynamically set for different regions and different meteorological elements for rapid model training, and the problem of adaptability of fixed time windows in a traditional method is solved.

The embodiment of the invention discloses a service deployment method and device, electronic equipment and a storage medium, and can solve the problems that when a dependency relationship is configured, the whole document needs to be set or modified, the complete dependency relationship among multiple services cannot be visually reflected, and the visibility is poor. Obtaining initial dependency information, wherein the initial dependency information is used for describing a dependency relationship between system service nodes; according to the initial dependency information, an initial dependency graph is constructed, the initial dependency graph comprises a dependency relationship network of at least one version, and the dependency relationship network comprises a plurality of system service nodes and at least one dependency relationship existing between the system service nodes; performing data configuration on the dependency relationship network of at least one version through pre-stored target configuration data to obtain a target dependency model, the target configuration data comprising configuration data for the plurality of system service nodes and configuration data for at least one dependency relationship; in response to the deployment instruction, the target dependency model is deployed.

The invention discloses an electric power system evaluation method based on wind power decision uncertainty, and relates to the technical field of electric power system reliability evaluation, and the method comprises the steps: obtaining the basic parameters of an electric power system, and carrying out the modeling of the reserve capacity of a wind driven generator according to the basic parameters; the reserve capacity is quantified based on the adjustable range of the pitch angle of the wind generating set, and a decision dependence model of the wind generating set is established according to the supply of the reserve capacity; and establishing a power system reliability evaluation method according to the decision relation between the wind speed prediction error and the pitch angle of the wind driven generator, and establishing a Markov model to monitor the randomness characteristics of power system elements. According to the method, the challenge on the uncertainty of the wind power system in a traditional evaluation method is effectively solved, the accuracy and the reliability of the operation state of the wind power system are improved, and the adaptability and the stability of a power system are improved. The operation state of the wind power system is monitored and evaluated in real time, and powerful support is provided for operation, maintenance and management of a power system.

The invention belongs to the technical field of power system steady-state standby and dynamic support resource cooperative scheduling, and provides a power system steady-state-dynamic resource scheduling method considering frequency modulation capability dependency, which comprises the following steps: constructing a data-driven multi-wind power plant frequency modulation capability and system decision variable dependency high-dimensional linear mapping model; based on the high-dimensional linear mapping model, constructing a system steady-state operation and frequency support characteristic integrated optimal scheduling model considering a frequency modulation capability dependent characteristic; and solving the integrated optimal scheduling model by adopting a two-stage decomposition solving strategy. According to the method, the frequency modulation capability dependent model is embedded into the system operation mode and frequency support characteristic integrated optimization framework, and a master-slave decomposition solving strategy is adopted, so that the calculation complexity is effectively reduced, and the efficiency of the system is improved on the premise that the system frequency safety and the fan operation safety are guaranteed. And economic optimal configuration of steady-state energy reserve and frequency support reserve of the system is realized.

The invention provides an operation and maintenance technology service problem prediction method and system based on big data analysis, and relates to the technical field of operation and maintenance service, and the method comprises the steps: firstly constructing an operation and maintenance service link dependency model containing a dependency relationship type, tracking the operation state of a link unit in real time, and recognizing an initial deviation unit; tracking the deviation transmission path to form an operation and maintenance service deviation transmission track, performing scene mapping on the operation and maintenance service deviation transmission track and a historical case library, and extracting matched problem types and development characteristics; and finally, processing big data based on a mapping result and a historical operation and maintenance problem, generating a prediction result containing a problem type, an affected link unit range and a problem development direction, and associating an operation and maintenance technology service problem intervention scheme set. According to the invention, the accuracy and timeliness of operation and maintenance problem prediction are improved.

The invention relates to a simulation calculation method for the flow velocity in the pipeline purging process, and the method comprises the steps: calculating the first density and the second density of a fluid medium for purging through employing an ideal gas model and a constant model, and calculating the first viscosity and the second viscosity through a strong temperature dependence model and the constant model; and performing weighted summation on the first density and the second density to obtain corrected density, performing weighted summation on the first viscosity and the second viscosity to obtain corrected viscosity, comparing the calculated flow velocity obtained by performing fluid field transient solution based on the corrected parameters with the actual flow velocity, and adjusting the weight through iteration until the error meets the requirement. And finally, when the error meets the requirement, performing flow field solution and particle motion tracking on the obtained effective density and effective viscosity to obtain a simulation result. According to the method, a dynamic coupling calculation model and weight iteration adjustment are combined, multi-dimensional parameters such as temperature, pressure and medium characteristics are integrated, and the method can be suitable for complex working conditions including different medium characteristics and changeable environmental conditions.

The application provides a cloud configuration item intelligent optimization system and method, and relates to the technical field of cloud configuration data optimization. The system comprises the following steps: mapping configuration item data to different levels of a rule dependency model to construct a dynamic configuration relationship graph, generating a multi-modal evidence body and a dynamic trust chain rule based on the dynamic configuration relationship graph and key attributes of the configuration item, and constructing a cause-effect dependency graph through instantiation, calculating a robustness index of the cause-effect dependency graph, performing conflict resolution arbitration, and determining the state of the configuration item according to the arbitration result; improving data accuracy and self-healing ability through automatic conflict resolution, enhancing the real-time performance and consistency of configuration optimization through cause-effect dependency analysis, and improving optimization efficiency while ensuring reliability.

The application discloses a kind of crane track joint real-time detection method and system, belong to crane technical field, by obtaining the joint image under a variety of different track surface conditions and extracting region of interest, combine model activation consistency index and background disturbance sensitivity, calculate background dependence index, and the dependence classification of model is carried out;Further, the recognition accuracy of high dependence model under new environment is predicted based on track background characteristics, and the prediction result is combined with dependence analysis, and the mechanism such as attention guided anti-correlation training is used to directional optimization of model, the method effectively overcomes the problem that the recognition ability of traditional CNN model decreases under the background change environment, significantly improves the generalization ability and deployment robustness of model, reduces the environment migration cost and misjudgment risk.

This application provides a method and system for predicting operational and maintenance (O&M) technical service issues based on big data analysis, relating to the field of O&M service technology. First, it constructs an O&M service link dependency model including dependency relationship types, tracks the operational status of link units in real time, and identifies initial deviation units. Next, it tracks the deviation propagation path to form an O&M service deviation propagation trajectory, maps it to a historical case database in a scenario-based manner, and extracts matching problem types and development characteristics. Finally, based on the mapping results and historical O&M problem processing big data, it generates prediction results including problem type, affected link unit range, and problem development direction, and associates them with a set of O&M technical service problem intervention schemes. This invention improves the accuracy and timeliness of O&M problem prediction.

The invention discloses an intelligent concurrency control method and system for large language model reasoning, and the method aims at a to-be-deployed model and a reasoning scene, enumerates a feasible configuration combination, and achieves the prediction of the throughput, first token delay, token generation delay and other performance indexes based on a single-target or multi-target regression model. The single-target optimization selects the optimal performance configuration, and the multi-target optimization adopts a Pareto optimization method to obtain a trade-off solution set, so that the automatic configuration recommendation of the concurrency control parameters is realized. The method does not need to depend on internal structure parameters of the model and trial and error adjustment parameters, does not need to perform explicit analysis on calculation load and communication overhead of each layer of Transform, has good universality, and can effectively improve performance and resource utilization efficiency of large language model reasoning service.

The application provides a kind of based on the method and device of chain analysis of propagation, electronic equipment and storage medium, related to distributed storage technical field, by obtaining target system historical operation data and extracting time series fault event set, identify the causal propagation relationship between different fault event types containing propagation direction, probability and time delay, build inspection dependency model, formulate and execute the inspection strategy containing priority scheduling and abnormal traceability evaluation, and can be based on new data dynamic update optimization related knowledge and model, therefore, it can solve the problem of low efficiency of inspection, difficulty in root cause positioning and inability to effectively contain fault propagation in the prior art due to isolated inspection items, unquantified fault propagation characteristics and rigid inspection strategies, achieve the technical effects of intelligent and accurate inspection, significantly improving the efficiency of inspection, quickly locating the root cause of the fault, preventing fault propagation in advance, and continuously optimizing the model to ensure long-term stable and efficient operation.

The invention provides an attribute dependency relationship-based H5 page generation method and computer equipment, and the method comprises the steps: defining a page structure through a structured description file, instantiating a component and binding attributes, constructing an attribute dependency model to realize attribute linkage updating, responding to a layout, adjusting a dynamic calculation layout, and exporting an optimization project file. The problem that a page generated by an existing H5 page generation scheme cannot adapt to display requirements of different terminals is solved.