30 results about "Probability assignment" patented technology

The invention discloses a time-frequency domain evidence fusion harmonic variable working condition detection method and device suitable for non-stationary harmonic data, electronic equipment and a storage medium, and belongs to the technical field of electric digital data processing. Performing fast Fourier transform on the original current waveform time domain data; dividing original current waveform time domain data into subsequences, and calculating a nonlinear weighted standardized Euclidean distance; calculating the posterior probability distribution of the time interval of the point change moment, and recording the number of times that each moment is deduced as a point change; carrying out normalization processing on the contour of the nonlinear weighting matrix and the number of times that each moment is deduced as a change point, respectively obtaining a sequence arranged according to the moment, intercepting a sub-sequence, calculating shape similarity, and obtaining an interval in which the change point exists; selecting candidate moments exceeding a threshold value as a mutually exclusive recognition framework, and constructing and synthesizing a basic probability distribution function; and calculating a trust function and a likelihood function, and obtaining a final change point through a discriminant formula. According to the method, the change point can be accurately and effectively found.

The invention provides a grid data processing method, system and device of a three-dimensional virtual model and a medium, and belongs to the technical field of three-dimensional virtual model processing, and the method specifically comprises the following steps: receiving original grid data; performing feature extraction and probability distribution on the vertex set; constructing an expansion graph through a k-nearest neighbor algorithm, calculating an edge connection probability, and multiplying the original adjacency matrix by the attention weight matrix to generate a simplified adjacency matrix; and performing feature coding and probability classification on the candidate triangle set, filtering and correcting non-manifold edges to obtain a simplified triangle set, and transmitting the simplified triangle set to target equipment. According to the invention, the lightweight processing of the original grid data is realized. On the premise that key geometric and topological characteristics of the model are not affected, the operation efficiency of the model on the target equipment is improved, the loading time is shortened, the display fluency is improved, and powerful support is provided for application of the three-dimensional model in scenes with high model response speed requirements such as virtual assembly, analogue simulation and real-time rendering.

The invention relates to the technical field of named entity recognition, in particular to a named entity recognition method and system based on long-distance information enhancement and boundary smoothing. Encoding an input sentence X by using a pre-trained Chine BERT encoder, and generating context representation information hi; carrying out feature extraction on the corpus comprising the context representation information by using a bidirectional long short-term memory network and a SwinTransform with a mask; the extracted features are sent to two independent multi-layer perceptron (MLP) to represent starting and ending of the entities respectively, and the span score of each entity is calculated; performing convolution on all entity span scores by using a convolutional neural network, and converting spatial information among words into spatial features; assigning a small part of probability Epsilon to a span adjacent to the current label, assigning the remaining probability 1-Epsilon to the span of the current label, performing boundary smoothing operation on each label to obtain a smoothed label # imgabs0 #

The invention relates to the technical field of product reliability evaluation, in particular to a product reliability evaluation method and system based on multi-information measurement fusion. The method comprises the following steps: acquiring an unknown parameter of life distribution adapted to an evidence source, generating a parameter interval for the unknown parameter, and generating basic probability distribution associated with the parameter interval based on the evidence source; calculating correlation coefficients between the evidence sources based on the basic probability distribution, and generating importance weights according to a plurality of correlation coefficients of the evidence sources and other evidence sources; calculating the information entropy of the evidence source, and obtaining a credibility weight through normalization based on the information entropy; carrying out normalization on an operation result of the importance weight and the credibility weight to obtain an evidence weight; taking the evidence weight as a weight coefficient of basic probability distribution, and obtaining comprehensive probability distribution through evidence fusion; and carrying out parameter estimation on unknown parameters based on comprehensive probability distribution, and substituting an estimation result into a reliability function to carry out reliability evaluation, so that reliability evaluation under fault-free data is realized.

The invention discloses a decision-level-oriented scene information inconsistency elimination system and a decision-level-oriented scene information inconsistency elimination method, and relates to the technical field of artificial intelligence. Comprising a human body behavior data acquisition module, a human body behavior data preprocessing module, a bidirectional long-short-term memory human body behavior recognition network module, an inconsistency detection module, an initial basic probability distribution mapping module, an evidence set conflict degree division module, a self-adaptive credibility evaluation module and a human body behavior information application module which are connected in sequence. The human body behavior data acquisition module comprises an inertial motion sensing unit and a physiological signal monitoring unit. The technical problem to be solved by the invention is to provide a decision-level-oriented scene information inconsistency elimination system and method, which are used for realizing effective fusion of conflict evidences, fully utilizing complementarity of multi-source information and improving reliability and accuracy of decisions in complex scenes.

An automated functional fault localization method for software product lines integrates techniques such as Spectrum-Based Fault Localization (SBFL), machine learning algorithms, data mining, and information theory. By utilizing code blocks as the detection granularity, it performs probability allocation from three perspectives: prediction, actual execution, and correlation. The suspiciousness values of code blocks are then computed using uncertainty reasoning algorithms, enabling the rapid identification of suspicious statements. Building on this, a more precise assessment of these statements is achieved by evaluating at three levels of granularity: global, local, and code block. This ultimately results in more efficient and accurate fault localization. The code blocks directly correspond to internal feature interactions, significantly enhancing the efficiency of searching for these interactions. For program statements with a higher likelihood of causing software errors, this method allows for quick identification and localization without manual intervention, thereby facilitating the maintenance of software product line systems.

The invention provides a task allocation method and device, equipment, a storage medium and a program product, relates to the technical field of data processing, and can dynamically adjust a task scheduling strategy according to the task load and the state of a proxy object so as to improve the resource utilization efficiency. The method comprises the following steps: obtaining the total quantity of tasks and the quantity of proxy objects, and determining the performance index of each proxy object; and based on the total task quantity and the proxy quantity, a plurality of resource allocation schemes are initialized and generated, and each scheme allocates different task quantities to the proxy objects. And calculating the fitness score of each scheme according to the performance index, and selecting the scheme with the highest score as a target resource allocation scheme. And according to the performance index of each proxy object, determining the scheduling weight of the proxy object in the target scheme. And for each to-be-scheduled task, calculating a selection probability based on the scheduling weight of each current proxy object, and allocating the proxy objects according to the probability. And after distribution, updating the scheduling weight of the proxy object, and repeatedly executing the process until all tasks are distributed.

The application discloses a content detection method and device, equipment, a storage medium and a program product, and relates to the technical field of artificial intelligence. The method comprises the following steps: acquiring user input content to be detected; detecting the user input content by at least two target detection engines in M detection engines, obtaining a determination detection result output by each target detection engine and a first confidence of the determination detection result; determining probability distribution information of detection results of different categories of each target detection engine according to the confidence of each target detection engine and the first confidence of the determination detection result; determining target confidence of the detection results of different categories of the at least two target detection engines according to the probability distribution information and a conflict factor; and the target confidence is used for determining a detection result of the user input content. In this way, the problem that the content detection result is inaccurate can be solved.

Disclosed herein are system, method, and computer program product embodiments for determining a probability of a device at risk. The device may be associated with a plurality of security parameters. For a security parameter, the device can be in multiple states. A probability value corresponding to a security parameter can indicate the security parameter being in a state among the multiple states. A probabilistic graphical model may be used to represent dependences of the plurality of security parameters. A device security risk prediction module may determine a probability of the device at risk based on the probabilistic graphical model and the probability assignments to the plurality of nodes of the probabilistic graphical model, and further determine a user action instruction to be provided to a user of the device based on the probability of the device at risk.

The invention discloses a process parameter decision-making system and method based on multi-source information fusion, and relates to the technical field of conflict detection.The process parameter decision-making method comprises the steps that multi-source heterogeneous data in the process production process is collected, and feature extraction and uncertainty quantification are conducted on the preprocessed multi-source heterogeneous data; constructing a process knowledge graph, calculating distances and conflict entropies among different evidences, calculating evidence credibility weights by using the distances among the evidences, and correcting each evidence by using the credibility weights; carrying out evidence fusion by using an evidence theory to obtain a probability distribution function of comprehensive evidence; determining a current process state and confidence according to the fused evidence, extracting different process parameter adjustment schemes associated with the current process state from the knowledge graph, and selecting an optimal process parameter adjustment scheme by using a neural network algorithm; and executing the optimal process parameter adjustment scheme, collecting multi-source heterogeneous data of the next production cycle, and judging whether the optimal process parameter adjustment scheme is effective or not.

The invention belongs to the field of hospital scheduling, and relates to a dynamic probability allocation mechanism-based transporter scheduling method, which comprises the following steps of: extracting historical transportation data to obtain a transportation time matrix; obtaining a patient transportation matrix based on the current patient transportation data; processing the transporter data based on the scheduling tendency of the patient to obtain an initial transporter scheduling matrix; the scheduling tendency is related to transporter selection of the first transportation task and the second transportation task; the transporter data comprises the workload and the current position of a transporter; calculating scheduling information of transporters in the initial transporter scheduling matrix, and dynamically updating the transporter scheduling matrix through a dynamic transition probability model based on the scheduling information to obtain a final transporter scheduling matrix; the objective of the invention is to solve the pain point problems of unbalanced workload of central transporters, long waiting time of patient examination and the like under traditional experience scheduling, and realize the normal form transformation of central transporter scheduling management from experience driving to data driving.

The embodiment of the invention relates to the technical field of computers, and provides a method and device for randomly distributing users, equipment, a storage medium and a product, and the method comprises the steps: carrying out the hash processing of data of a plurality of preset random sources, and obtaining an initial random seed value; generating a random number sequence by using a random number generation model according to the initial random seed value and the number of users to be distributed; wherein the random number sequence comprises a random number corresponding to each user to be distributed; calculating a probability allocation interval corresponding to each target domain according to a preset user allocation proportion of the plurality of target domains; and allocating each user to be allocated to a corresponding target domain according to the probability allocation interval and the random number sequence. According to the embodiment of the invention, the reliability and safety of random distribution of the user can be improved.

This invention discloses a decision-level contextual information inconsistency elimination system and method, relating to the field of artificial intelligence technology. Its features include: a human behavior data acquisition module, a human behavior data preprocessing module, a bidirectional long short-term memory human behavior recognition network module, an inconsistency detection module, an initial basic probability allocation mapping module, an evidence set conflict degree division module, an adaptive credibility assessment module, and a human behavior information application module, connected sequentially. The human behavior data acquisition module includes an inertial motion sensing unit and a physiological signal monitoring unit. The technical problem this invention aims to solve is to provide a decision-level contextual information inconsistency elimination system and method to achieve effective fusion of conflicting evidence, fully utilize the complementarity of multi-source information, and improve the reliability and accuracy of decision-making in complex scenarios.

The invention provides a reinforcement learning-oriented multi-track cross-state award attribution method, and belongs to the field of reinforcement learning and graph neural networks. According to the method, in each training process, track data generated by environment interaction are collected by adopting a traditional reinforcement learning method, and strategy loss and value function loss are calculated according to an original mode; meanwhile, a global state transition diagram is constructed based on the trajectory data, state attribution probability distribution is calculated by using a graph neural network, each trajectory final reward is allocated to a state node according to an attribution probability, and a reward attribution signal of each state node is obtained; a loss function, a value function, and a reward signal are modified based on the obtained reward attribution signal. Through the integration mode, the method not only can make full use of the advantages of information sharing and cross-state reward attribution of multiple tracks, but also can be seamlessly embedded into a mainstream reinforcement learning training framework, and can significantly improve the training effect and strategy performance of reinforcement learning in a sparse and delayed reward environment.

The invention discloses a training method of a file fragment classification model and a file fragment classification method, and relates to the field of file fragment classification, and the method comprises the steps: collecting different types of original file data, and carrying out the random sampling of data blocks of the different types of original file data at a fixed byte length; converting each data block into a corresponding initial token sequence; converting each initial token sequence into a target token sequence; and pre-training a Transform model based on each target token sequence after dynamic masking, the dynamic masking comprising: randomly selecting byte tokens with a preset proportion in the target token sequence as target byte tokens, and performing different masking operations on each target byte token based on probability allocation, the different mask operations comprise replacing the target byte token with the fourth function token, randomly replacing the target byte token with other byte tokens and keeping the target byte token unchanged. According to the method and the device, the generalization ability of the Transform model can be improved.

This invention discloses an event-driven emergency scenario simulation method for reservoir dam emergencies. It constructs an ontology model of the reservoir dam emergency scenario based on event ontology theory; generates a multi-entity Bayesian network model according to the mapping rules between the ontology model and multi-entity Bayesian segments; removes irrelevant branches to construct a specific scenario Bayesian network topology; selects a probability assignment method based on the logical relationships and mechanisms between nodes, initializes probability parameters, and completes the construction of the emergency scenario simulation model; collects current scenario information as evidence and assigns probability values ​​to the root node; uses the driving event as an input variable and determines the probability values ​​of the state variables and output variables of subsequent scenario nodes according to inference rules; and achieves probabilistic simulation and emergency decision assessment of reservoir dam emergency scenarios. This invention enables accurate prediction of reservoir dam emergency scenarios.

A method for measuring the importance of social network nodes based on multi-feature fusion uses centrality indicators, transitivity indicators, and prestige indicators as attribute features for measuring the importance of nodes; secondly, it fuses the three features. In this process, a probability assignment generation method based on the interval number model is used to transform the feature values into basic probability assignment functions, and then the improved evidence theory is used to fuse multiple BPAs; finally, the probability of node importance is obtained based on the fused indicators, and the nodes are sorted. When measuring the importance of nodes in a real large-scale social network, this method has high measurement accuracy.

This invention discloses a time-frequency domain evidence fusion method, apparatus, electronic device, and storage medium for detecting harmonic variations in non-stationary harmonic data, belonging to the field of electrical digital data processing technology. The method involves performing a Fast Fourier Transform on the original current waveform time-domain data; dividing the original current waveform time-domain data into subsequences and calculating the nonlinear weighted normalized Euclidean distance; calculating the posterior probability distribution of the time intervals between change points and recording the number of times each time point is inferred to be a change point; normalizing the nonlinear weighted matrix contour and the number of times each time point is inferred to be a change point to obtain sequences arranged by time; truncating subsequences and calculating shape similarity to obtain the intervals where change points exist; selecting candidate times exceeding a threshold as mutually exclusive identification frames and constructing a synthetic basic probability allocation function; calculating the trust function and likelihood function, and obtaining the final change point using a discriminant formula. This invention can accurately and effectively find change points.

A method may store time series data that includes a biophysical response over sequential time periods. An initial variable can be established having an event value corresponding to each time period. A plurality of assigned variables can be generated, each having an assigned event value corresponding to each time period with one assigned event value being different with respect those of the initial variable and the other assigned variables. The initial and assigned variables can be evaluated with a probability function to determine the variable having a highest probability of event occurrences with respect to the biophysical responses. Using the highest probability initial or assigned variable as the initial variable, generation of assigned variables and a highest probability determination can be repeated until a highest probability variable has been determined. The highest probability variable can be used to predict the biophysical response in a user. Corresponding systems are also disclosed.

The Beidou PPP-RTK multiple risk source trusted probability allocation method and device relate to the field of trusted probability allocation. In order to solve the problem in the prior art that the trusted probability allocation method based on the traditional civil aviation field is no longer applicable due to the large difference between the technical means of the Beidou PPP-RTK carrier phase layer and the navigation and positioning method of the civil aviation pseudorange layer, and trusted monitoring is therefore difficult to carry out, the technical solution provided by the present invention is: Beidou PPP-RTK multiple risk source trusted probability allocation method, the method includes: constructing a Beidou PPP-RTK trusted monitoring fault tree model; based on the model, obtaining an error distribution model; based on the model, constructing a protection level equation with multiple hypothesis solutions separated; constructing an optimal allocation model; solving the optimal allocation model to obtain an allocation result; constructing a protection level to determine whether a stable state has been reached; and outputting the final allocation result. It is suitable for the research work on the trusted probability allocation method for Beidou PPP-RTK, and is also suitable for ensuring the credibility of Beidou PPP-RTK navigation and positioning services.

The application relates to the field of automatic driving, and particularly discloses a method for predicting takeover success probability and evaluating intervention effect under human-machine co-driving conditions. The method comprises the following steps: step 1, establishing a takeover success sample: determining the optimal takeover budget time and the takeover success rate cumulative curve of a driver, and constructing a mapping parameter set; step 2, predicting the takeover fitness of the driver: optimizing parameters and constructing a real-time takeover state vector T, calculating the Mahalanobis distance D M (S) of the real-time takeover state vector and a perfect takeover state vector D M (F), calculating the basic trust probability assignment m(s) of perfect takeover and the basic trust probability assignment m(f) of a failed takeover, and setting the decision threshold of the trust degree assignment as epsilon, and predicting that the driver can successfully take over when m(s)-m(f) > epsilon. The method for predicting takeover success probability and evaluating intervention effect under human-machine co-driving conditions can improve prediction accuracy and efficiency, and reduce takeover risks.

The invention discloses a high energy efficiency WSN topology control method based on SP-DPC clustering. The method is based on an SP-DPC clustering algorithm, SP-DPC is an algorithm based on density peak clustering, the problem that DPC-MND is poor in performance when facing manifold cluster data sets is solved, meanwhile, the thought of transmission probability allocation is introduced, and associated allocation errors when non-cluster-head nodes are classified into clusters are prevented. Meanwhile, during cluster head election, the distance from the elected node to the base station is used as a dynamic factor, and the problem of energy holes around the base station is solved. In the routing path design stage, unified calculation is carried out by the base station. A path weight matrix is constructed by integrating factors such as node energy and distance, and an optimal path is selected through a Dijkstra algorithm. Compared with other protocols, the protocol provided by the invention has remarkable performance improvement in the aspect of energy efficiency.

The invention discloses a day-ahead multi-time scale probability scheduling method for a power system based on joint chance constraint. The method comprises the following steps: 1, acquiring day-ahead predicted output data of a renewable power supply and a probability distribution parameter of output of the renewable power supply; 2, establishing a multi-time scale electric power system day-ahead scheduling model, and introducing a joint opportunity constraint into the model to carry out risk control on the system; 3, equally dividing the total risk probability, carrying out probability allocation on a single opportunity constraint, and solving an optimization problem to obtain an initial solution; 4, iteratively adjusting opportunity constraint probability distribution by adopting a projection gradient method so as to reduce the conservative property of a solution; 5, solving the optimization problem again based on the updated opportunity constraint probability distribution scheme, and evaluating the overall risk of the system; and 6, if the system risk level converges to a target value, outputting a final scheduling scheme, otherwise, continuing iteration. According to the method, the optimal allocation of the chance constraint risk probability can be realized, so that a low-conservative scheduling scheme meeting the system risk level can be obtained.

The present invention belongs to the field of hospital scheduling and relates to a transporter scheduling method based on a dynamic probability allocation mechanism, comprising: extracting historical transport data to obtain a transport time matrix; obtaining a patient transport matrix based on current patient transport data; processing the transporter data based on the patient's scheduling tendency to obtain an initial transporter scheduling matrix; the scheduling tendency is related to the transporter selection for the first transport task and the second transport task; the transporter data includes the transporter's workload and current location; calculating the transporter's scheduling information in the initial transporter scheduling matrix, and dynamically updating the transporter scheduling matrix based on the scheduling information through a dynamic transfer probability model to obtain a final transporter scheduling matrix; the method aims to solve the pain points of unbalanced workload of central transporters and long waiting time for patient examinations under traditional experience-based scheduling, and realize the transformation of central transporter scheduling management from an experience-driven to a data-driven paradigm.

Systems and methods for progressively releasing features used in conjunction with a service in a computing environment are disclosed. The computing environment includes users and resources. A method includes providing access to the service to the plurality of users and providing a particular feature to be used in conjunction with the service. The method can additionally include maintaining a directory of the entities of the computing environment. The entities can include the users and / or the resources. The directory of the entities organizes the entities into groups. The example methods can also include assigning an access value associated with the particular feature to a particular group of the groups. The access value can indicates an assigned probability that the feature will be enabled with respect to a particular entity included in the particular group. The assigned probability can advantageously be greater than zero and less than one.

The invention discloses a fault line selection method and device based on multi-criterion fusion, equipment, a medium and a product. The method comprises the steps of determining a first fault measurement value of each line under at least two fault criteria when it is determined that the power distribution network has a fault; according to the difference between each first fault measurement value and each fault criterion, determining a second fault measurement value of each line under different fault criteria; determining a third fault measurement value of each line under different fault criteria according to the difference condition of the second fault measurement values between the lines; performing multi-criterion fusion based on the basic probability distribution function and the D-S evidence theory, and determining a corresponding fault line selection result according to the fault probability value of each line after the multi-criterion fusion; wherein the basic probability distribution function is constructed based on each third fault measurement value. Fault measurement can be effectively processed by using transverse and longitudinal relations between fault criteria and between lines, and reliable line selection of power distribution network faults is realized based on multi-criterion fusion.

This invention discloses a keyword generation method based on category guidance and alternating meta-tuning. Addressing the probability imbalance caused by diverse domain distributions, the method first constructs a contrastive learning retrieval system that integrates category information, and then selects a high-quality support set to construct a pseudo-task. Next, the generation model parameters are decoupled into a high-sensitivity weighted sub-network and a backbone network. Then, alternating meta-tuning is performed: a three-step loop of simultaneous updates to both networks, updating the weighted sub-networks separately with a fixed backbone, and updating the backbone using new weights. Finally, a smoothing factor is introduced to update the retrieval system in reverse. This invention effectively calibrates the probability allocation between replication and generation through parameter decoupling and category guidance, improving the generation accuracy of keywords in both present and missing categories, as well as the model's generalization ability.

This invention relates to a method, system, and medium for pipeline routing on offshore platforms based on ant colony optimization algorithm, belonging to the field of intelligent optimization algorithm technology. The method includes the following steps: S11: Environment Modeling: Discretizing the three-dimensional space of the offshore platform into a mesh model, marking obstacle areas and feasible areas, and calculating the energy values ​​of mesh nodes; S12: Parameter Initialization: Constructing a heuristic information matrix based on pipeline inlet / outlet coordinates and equipment parameters. The matrix determines the node transition probability through a direction probability allocation strategy and a distance-sensitive function; S13: Path Optimization: Generating candidate paths through dynamic smoothing search, local pheromone adaptive adjustment, and collaborative niche search, and updating the pheromone; S14: Constraint Verification and Result Output: Verifying whether the path meets engineering constraints, and outputting the three-dimensional coordinates, length, and number of bends of the optimal pipeline path after iterative optimization. The method of this invention achieves efficient, safe, and economical pipeline routing on offshore platforms.

The invention provides a DS theory-based recognition method and system for multiple emotion recognition models, and the method comprises the steps: obtaining a comment emotion data set and a dialogue data set, and carrying out the preprocessing of the obtained data; constructing and training a plurality of independent emotion recognition sub-models by utilizing the preprocessed data set, and respectively outputting probabilities; a weight coefficient alpha i of each sub-model is distributed, an uncertainty coefficient 1-alpha i is introduced, the sub-model output probability is combined, basic probability distribution is completed, and a quality function is constructed; combining the quality functions of the four sub-models into a new quality function; calculating a sub-trust function based on the new quality function, and then calculating a combined likelihood function; the trust function and the likelihood function form a trust interval which represents the confirmation degree of emotion recognition, and final probability distribution is determined according to needs; and performing emotion judgment based on probability distribution.

The invention relates to a computer-implemented method for generating training data with at least one foundation model and for training based on the generated training data. The method comprises providing (S1) a first training data generation model; providing (S2) a class distribution model; providing (S3) a classification model; providing (S4) a second training data generation model; training (S5) the first training data generation model as a function of a class condition, including the class distribution model; generating (S6) first training data as a function of the class condition, including real data, by the first training data generation model;training (S7) the classification model based on the first training data, such that at least one class and / or a class probability is assigned to the respective first training data by the classification model, and second training data classified in this way is generated; and training (S8) the second training data generation model based on the first training data, including real data;