39 results about "Probabilistic inference" patented technology

The present invention discloses a system and method for secure artificial intelligence-based financial technology governance and risk management, designed to provide real-time, autonomous, and verifiable compliance assurance within digital financial ecosystems. The invention integrates a secure artificial intelligence processing unit, a governance control processor, a cryptographically anchored storage unit, a federated learning coordination processor, and a quantum-resistant communication interface enclosed within a tamper-proof hardware structure. The system performs encrypted machine learning computations on financial transaction data using homomorphic encryption and trusted execution environments to preserve confidentiality during analysis. It computes a governance risk index based on probabilistic inference and anomaly detection to identify regulatory deviations, applies adaptive compliance reasoning across multi-jurisdictional frameworks, and automatically enforces governance actions through secure decision logic.

The invention discloses an unmanned aerial vehicle positioning method based on a Bayesian network, and the method comprises the steps: collecting unmanned aerial vehicle data, carrying out the preprocessing, obtaining sound wave data, motion data, infrared image data and environment data, carrying out the spectrum analysis based on the sound wave data, obtaining sound spectrum time sequence data, and carrying out the positioning of the unmanned aerial vehicle. Constructing a probability density thermodynamic diagram and extracting a first range through a sparse variational Gaussian process according to time, frequency and energy three-dimensional information of the sound spectrum time sequence data and environment data, and obtaining first positioning data through a Bayesian network according to the first range and motion data, and performing adaptive wavelet down-sampling on the infrared image data and the sound spectrum time sequence data in the lateral direction to obtain lateral positioning data, and performing deviation correction on the first positioning data by using the lateral positioning data to obtain second positioning. According to the method, through space-time synchronization, adaptive processing and probabilistic reasoning of sound spectrum time sequence data, the positioning precision and stability in a complex environment are improved.

The invention discloses an industrial product quality risk association method based on a graph neural network, and relates to the field of industrial manufacturing, and the method comprises the steps: collecting multi-source industrial data, carrying out the preprocessing of the multi-source industrial data, obtaining a standardized data table, and constructing an industrial element heterogeneous graph based on the standardized data table; inputting the industrial element heterogeneous graph into a graph attention network basic model for node representation learning, and setting a supervised learning task taking a product batch node in the industrial element heterogeneous graph as a prediction target and taking a historical quality detection result as a supervised label; and introducing a meta-path self-learning layer in a message passing process of the graph attention network basic model. According to the method, deterministic prediction is upgraded into probabilistic inference, internal confidence degree measurement is provided for each risk prediction value, and automatic, accurate and interpretable correlation analysis and traceability positioning of the industrial quality risk are realized.

The invention discloses an unplanned secondary operation decision support method and system based on EMR analysis. According to the method, firstly, multi-source EMR data are collected in real time through an HL7 FHIR standardized interface, differential analysis is conducted through a time sequence analysis algorithm and a BioBERT model fused with a medical knowledge graph, and standardized clinical indexes are generated; secondly, the indexes are matched with an unplanned secondary surgery thematic knowledge graph in real time, multi-path probabilistic reasoning is carried out based on a Bayesian network, and a risk conclusion with confidence and a reasoning chain are output; and generating a structured report including risk early warning, core basis, possible reasons and processing suggestions, and pushing the structured report in multiple channels. And finally, collecting clinical feedback, optimizing the knowledge graph and the analytical model, and forming a closed loop. According to the invention, dependence on experience of doctors is reduced, and accuracy and timeliness of risk research and judgment are improved.

The invention discloses a data center power distribution optimization method and system, and the method comprises the steps: building a three-dimensional environment-power consumption correlation field model through monitoring a data center calculation node power consumption signal and environment sensing data, and deeply mining an influence rule of environment factors, such as temperature and humidity, on a power consumption demand; probabilistic power consumption parameters are extracted from the association field by applying a probability inference analysis technology, a power consumption event triggering threshold value is established, and a dynamic power consumption management boundary is formed; a quality demand matrix is generated based on service quality demand difference identification, and a hierarchical power consumption parameter system is constructed; adaptive parameters are obtained through load-power consumption correlation analysis, a power consumption redundant component is identified and converted into a power consumption enhancement cache region, and self-adaptive power consumption adjustment is achieved; and finally, constructing a power consumption control matrix, detecting power consumption imbalance nodes, carrying out balanced pairing, and generating an accurate power consumption management instruction, thereby realizing intelligent optimization control of the power consumption of the data center, effectively improving the energy utilization efficiency and guaranteeing the service quality.

The invention relates to the technical field of model simulation, and discloses a precise traceability method and system for underground water organic pollutants, and the method comprises the steps: building a space-time feature set based on monitoring well concentration data and geological parameters, and generating a preliminary continuous concentration field through Gaussian process regression; secondly, extracting geometric features of a concentration field, establishing a flow field model in combination with an underground water level gradient, simulating a pollutant transport process, screening a potential source position set according with a time deviation threshold value, and constructing a source strong release model to invert optimal release parameters; a migration path is generated, and an effective path is screened according to the cumulative flux coverage degree to reconstruct a global concentration field; and finally, determining a final pollution source coordinate through concentration field gradient rotation analysis and isoline geometric continuity evaluation in combination with Bayesian probabilistic reasoning. According to the method, the traceability accuracy can be improved.

This invention discloses a method and system for controlling an automotive electronic vacuum pump based on braking intention recognition, belonging to the field of intention recognition. The method includes: acquiring vehicle status information, driver operation information, and environmental perception information from an advanced driver assistance module; calculating environmental risk levels based on environmental perception information and vehicle status information, including a forward collision risk level based on collision time and a downhill risk level based on road gradient; extracting current pedal feature parameters based on operation information, and inputting them along with the environmental risk level and historical braking tendency quantification values ​​into a Bayesian network to calculate the posterior probability distribution of braking intention at the current moment through probabilistic inference; generating pre-scheduled control commands for the electronic vacuum pump based on the posterior probability distribution of braking intention and a preset probability threshold, and sending them to the electronic vacuum pump controller for execution. This application solves the technical problem of existing automotive electronic vacuum pumps having slow response and difficulty in handling complex operating conditions, resulting in insufficient braking safety.

The present invention relates to a database (DB) inference attack control device and method and, more specifically, to a probabilistic logic-based database inference attack control device and method, in which: when receiving result data output from a database management system (DBMS) and inferring sensitive information from each result data record constituting the result data, data attributes are restricted on the basis of a sensitive inference attack element threshold, wherein the data attributes correspond to an inference attack element preconfigured to have a more critical impact compared to other inference attack elements; a risk probability of inferring the sensitive information is calculated on the basis of a preconfigured probabilistic inference attack logic; when the calculated risk probability exceeds a risk determination threshold, detection of an inference attack is determined, and one or more of data attributes corresponding to an inference attack element associated with the inference attack are restricted; and on the basis of an inference-possible logic, which is preconfigured as being usable for an inference attack, in addition to the probabilistic inference attack logic, data attributes corresponding to the inference attack element included in the inference-possible logic are restricted on the basis of an inference-possible logic threshold.

The invention provides a control method for accurate blanking of a hopper scale, and relates to the technical field of industrial automation and process control, and the method comprises the steps: obtaining the internal density distribution data of a material in a hopper through at least one electrical capacitance tomography sensor; obtaining inter-particle stress data of the material through at least one acoustic emission sensor; taking the internal density distribution data and the inter-particle stress data as observation variables, and inputting the observation variables into a dynamic Bayesian network DBN model; utilizing a DBN model to carry out probabilistic inference on a hidden material state of the material, the hidden material state at least comprising a flow failure probability; when the flow failure probability exceeds a preset threshold value, generating a preventive control signal; a preventative control signal is applied to a flow actuator to correct the flow condition of the material prior to a macroscopic flow failure of the material. The blocking risk can be reduced to the germination state as much as possible in the early stage of formation of the pre-bridging microstructure.

This application relates to the fields of artificial intelligence and computer technology, specifically providing a risk decision-making method and system for the entire lifecycle of power data. The method includes: acquiring multi-source heterogeneous log data from the entire lifecycle of a power information system; using an unsupervised feature extraction module based on a variational autoencoder to extract latent feature vectors and determine the VAE reconstruction error; performing spatiotemporal feature extraction on the latent feature vectors to obtain the temporal behavior probability from the time-series analysis stream and the spatial graph embedding distance from the spatial analysis stream; and constructing a dynamic decision-making module based on a Bayesian network, using the VAE reconstruction error, temporal behavior probability, and spatial graph embedding distance as multi-source evidence nodes for the dynamic decision-making module to calculate the posterior probability of risk events. This application, through the collaborative fusion of spatiotemporal features and causal probabilistic inference, can significantly improve the safety risk perception and dynamic control capabilities of power data throughout the entire process of acquisition, transmission, and processing.

The application discloses an industrial product quality risk correlation method based on a graph neural network, relates to the field of industrial manufacturing, and comprises the following steps: collecting multi-source industrial data, preprocessing the multi-source industrial data, obtaining a standardized data table, constructing an industrial element heterogeneous graph based on the standardized data table, inputting the industrial element heterogeneous graph into a graph attention network base model for node representation learning, setting a product batch node in the industrial element heterogeneous graph as a prediction target, and setting a historical quality detection result as a supervised label of a supervised learning task, and introducing a meta-path self-learning layer in a message passing process of the graph attention network base model. The application upgrades deterministic prediction to probabilistic inference, provides an inherent confidence measure for each risk prediction value, and realizes automatic, accurate and interpretable correlation analysis and traceability positioning of industrial quality risks.

Provided are a shape model setting circuitry for setting a shape model of a target structure, a crack candidate plane in the shape model, and an observation plane of the shape model, an estimation model generator for generating an estimation model obtained from a numerical analysis of a structural analysis model by sequentially changing a boundary condition of the crack candidate plane in the structural analysis model generated from the shape model, and a crack state analyzer for estimating a position and a size of the crack by obtaining a distribution of load and displacement in the crack candidate plane at the same time by probabilistic inference through the application of an observation plane deformation vector indicating deformation of the observation plane obtained from measurement values, the estimation model, and a latent variable indicating presence or absence of the crack in the crack candidate plane.

This invention discloses an epidural anesthesia assistance method and system based on pressure curve characteristics, including a data acquisition and preprocessing module, an anatomical layer segmentation module, a multidimensional feature extraction module, a template matching and initial classification module, a Bayesian probabilistic inference and fusion decision-making module, a risk assessment and feedback output module, and an online adaptive update module. By utilizing real-time pressure monitoring, negative pressure recognition, and biometric frequency extraction technologies, it assists in determining the position of the epidural puncture needle, significantly reducing the rate of epidural puncture mispuncture and improving clinical safety. Through a Bayesian online learning update mechanism, the model parameters can be continuously iterated based on clinically labeled data, adapting to different operator habits and sensor drift, achieving long-term stable operation and individualized calibration. This invention is easy to integrate and has low hardware costs, facilitating large-scale clinical deployment.

This invention discloses a smart home central control system and method based on multimodal fusion. The control system includes a multimodal interaction interface layer, a device execution layer, and a central controller. The central controller is communicatively connected to both the multimodal interaction interface layer and the device execution layer. The central controller includes an instruction receiving and parsing unit, a context awareness unit, a multimodal fusion decision engine, a scheduling and execution unit, and a feedback unit. The instruction receiving and parsing unit receives and parses interactive instructions; the context awareness unit acquires various dynamic data to provide the context information required for decision-making; the multimodal fusion decision engine is trained based on a hierarchical hybrid decision architecture, which includes a rule engine layer, a probabilistic inference layer, and a reinforcement learning layer. This invention achieves a more accurate, robust, and personalized smart home control experience through a dedicated fusion algorithm, a hierarchical conflict arbitration strategy, hardware and software co-optimization, and an adaptive feedback mechanism.

This invention discloses a machine learning-based method for predicting the compressive strength of geopolymer-stabilized soil, belonging to the field of building material prediction technology. The method includes: constructing a Bayesian prediction model for geopolymer properties based on theoretical strength values, combining a physical constraint layer and a Bayesian probabilistic inference layer; training the Bayesian prediction model to generate a trained model; inputting the mix proportion parameter vector of the geopolymer-stabilized soil into the trained model to perform multiple Monte Carlo sampling predictions to obtain predicted strength values ​​and physical constraint strength values; and performing statistical analysis on the predicted strength values ​​to generate prediction confidence intervals. This invention, by generating prediction confidence intervals and physical contribution parameters, achieves a simultaneous characterization of the distribution characteristics of predicted strength and the degree of mechanistic influence, and realizes a unified expression of uncertainty quantification and mechanistic contribution within a machine learning framework.

The invention relates to the technical field of task scheduling, in particular to a store patrol task dynamic allocation method based on load balancing. Comprising the steps of generating a task vector through a shop patrol task processing request; generating a node load vector through the mobile terminal state data; constructing a directed attribute graph, calculating transfer parameters, generating a space transfer matrix, and constructing a feature tensor; inputting the feature tensor into an evaluation network for processing, and outputting a score matrix; outputting candidate sites through a search algorithm; calculating a load variance variable quantity and an out-of-limit probability according to the candidate sites; inputting the load variance variable quantity and the out-of-limit probability into a multi-objective optimization function for calculation, and outputting a distribution strategy; and analyzing the distribution strategy, and generating a scheduling instruction containing a node routing sequence. According to the method, chain structured fusion is carried out through probability deduction of high-dimensional data, search pruning of physical boundaries and multi-target game of global dimensions, and long-term stability and dynamic balance of cluster load distribution are facilitated.

A system and method for implementing artificial intelligence using deterministic wave interference and collapse logic derived from the Total Wave Modified Schrödinger Equation (TWMSE). An artificial agent is modeled as a system wavefunction interacting with one or more observer wavefunctions representing electromagnetic, gravitational, weak, or strong fields. A collapse function determines when interference exceeds a threshold, producing deterministic action or comprehension. Field parameters adapt through feedback to enable learning, residual interference forms resonant memory, and computation is performed directly on optical, electromagnetic, or neuromorphic hardware. The invention provides a unified framework—Total Wave Artificial Intelligence (TWAI)—that integrates action, understanding, learning, memory, and physical embodiment through field-based collapse rather than probabilistic inference.

This invention discloses a method for tracing the source of construction power supply accidents based on causal inference, comprising the following steps: collecting data on current, voltage, switch status, and equipment operation events at the construction site, and performing time synchronization processing to generate a construction power supply state sequence; dividing the state sequence into time slices according to preset intervals to form a time window data set; constructing a dynamic Bayesian network model based on the time window data, establishing time slice nodes and conditional probability connections; assigning conditional probabilities to nodes and determining accident constraint nodes; performing reverse probabilistic inference starting from the accident constraint nodes, calculating the conditional probabilities of upstream nodes and generating a propagation path set; sorting the path probabilities, determining the target tracing path, selecting the node with the highest probability in the path as the root cause node, and outputting the root cause node and its corresponding path. This invention achieves causal source tracing analysis of construction power supply accidents through a time-sliced ​​dynamic Bayesian network and reverse probabilistic inference.

The invention comprises: a shape model setting unit (21) for setting a shape model of an object structure (7) to which an external force is applied, a crack candidate surface (7fc) of a crack expected to be hidden in the shape model, and an observation surface (7fo) of the shape model; an inference model generating unit (22) for sequentially changing the boundary conditions of the crack candidate surface (7fc) in a structural analysis model produced based on the shape model, and generating an inference model obtained by numerically analyzing the structural analysis model; and a crack state analyzing unit (3) for applying an observation surface deformation vector representing the deformation of the observation surface (7fo) obtained by the actual measurement value of the observation surface (7fo), an inference model, and a potential variable representing whether there is a crack (9) in the crack candidate surface (7fc), and simultaneously obtaining the distribution of the load and displacement in the crack candidate surface (7fc) through probabilistic inference, thereby inferring the position and size of the crack (9).

The invention discloses a method and system for establishing a nuclear power plant accident risk assessment model, and relates to the technical field of nuclear power plant accident management. Comprising the following steps: constructing a topological structure of a Bayesian network based on a power plant probabilistic safety analysis model; converting a logic relation in the model into a probability relation of a Bayesian network; assigning probability parameters to all nodes in the Bayesian network to complete quantification of the model; and carrying out sensitivity analysis on the quantized model and carrying out simplified verification based on an analysis result. According to the method, a static probabilistic safety analysis model is converted into a dynamic Bayesian network, a probabilistic reasoning framework capable of fusing real-time monitoring data is constructed, and sensitivity analysis is adopted to optimize the model, so that the calculation efficiency is improved while the precision is ensured; therefore, the problem of dynamic risk assessment caused by limited, uncertain and missing parameters under a serious accident is solved, the accident process and the radioactive release consequence are quickly and accurately predicted, and the emergency response capability of the nuclear power plant is remarkably improved.

The present application relates to the technical field of image node classification, and especially relates to a semi-supervised node classification method based on a graph convolution network and a probabilistic inference model, solves the technical problems in the background art, the method utilizes an edge complete tree to divide a graph, and lets a message propagation mechanism of the graph convolution network interactively execute inside and between subgraphs, integrates the similarity between nodes into the graph convolution network, so that the node state matrix in each layer satisfies the similarity constraint, in order to perform node classification, the method utilizes a conditional random field to model the correlation between node labels, and deeply fuses it with the graph convolution network fusing the node similarity. The method uses a kind of local first-order approximation in frequency domain graph convolution to realize convolution architecture, can learn the information of graph on hidden layer, also uses fast approximate convolution, can quickly, scalablely complete semi-supervised classification task based on point, can efficiently semi-supervised learning on large-scale graph data.

The embodiment of the invention relates to the technical field of plant management, and discloses an unmanned aerial vehicle-based vegetation AI identification management method, which comprises the following steps of: identifying wide-area remote sensing data to obtain all abnormal sub-areas and window detection areas in a target detection area; executing a second acquisition task on all abnormal sub-regions and window detection regions to obtain corresponding fine remote sensing data; determining vegetation state information of the abnormal sub-region according to the fine remote sensing data, determining a canopy shielding relationship in the window detection region according to the fine remote sensing data, and identifying and determining a potential under-forest region under canopy coverage; probabilistic inference is carried out on the vegetation state of the potential under-forest region through fusion of multi-source evidence, and an under-forest vegetation probability distribution diagram containing inferred vegetation categories and corresponding confidence coefficients is generated; and generating a comprehensive vegetation state report of the target detection area. And the overall identification management efficiency is improved through a layered acquisition mode.

The present application relates to resource scheduling technical field, specifically to the innovative resource self-organizing scheduling method based on ant colony cooperation and bayesian learning, including the following steps: S1, factor space state variable initialization; S2, structured reading of public clues; S3, probabilistic inference of opportunity credibility.The present application combines the decision-making and collaborative behavior of individual innovation subjects, uses dynamically updated public clues, opportunity credibility inference, subjective value evaluation and dynamic depiction of collaboration threshold, effectively optimizes resource allocation and collaboration mode, improves decision-making efficiency through bayesian inference in practical application, reduces the negative impact of information overload, and ensures efficient collaboration when the project scale expands through the introduction of synergistic effect and immediate adjustment mechanism, enhances the adaptability and continuous optimization ability of the system through the endogenous decay of feedback mechanism and environmental memory, thereby improving the utilization efficiency of innovative resources and the success rate of project execution.

The invention provides a model-based offset sensing agent offline reinforcement learning method and device. The method comprises the steps of obtaining an agent offline data set; training an integrated model according to the offline data set; replacing the environment where the intelligent agent is located according to the integrated model, and generating synthetic data; extracting samples from the offline data set and the synthetic data, and determining an offset perception reward according to a state transition classifier and a strategy classifier; wherein the state transition classifier is used for compensating difference between prediction of the integrated model and a real environment, and the strategy classifier is used for compensating distribution offset caused by strategy updating; and according to a sample corresponding to the offset perception reward, updating an intelligent agent operation strategy by adopting an SAC algorithm. According to the method, the problem of data distribution mismatching in a real environment and a model is solved by uniformly processing the offline data set and the synthetic data from a probabilistic reasoning framework.

Aspects of the present invention provide systems and methods for distributed signal processing of indoor localization signals wherein statistical algorithms and machine learning are used in place of a fingerprint map. The disclosure relates to calculation of angle and distance based on measurements of an indoor localization signal, followed by energy-efficient distribution of signal processing. Local signal processing is performed using any of multiple eigen structure algorithms or a linear probabilistic inference, before cloud-based signal processing is performed using a nonlinear probabilistic inference and machine learning that's been trained with historical data transmitted by the base stations and time-of-day location patterns. Without having to generate and constantly update an energy-exorbitant fingerprint map, the disclosed system reduces localization error to merely 50 cm with 95% probability without compromising energy-efficiency to rival the accuracy of indoor localization systems that utilize fingerprinting.

Aspects of the present invention provide systems and methods for distributed signal processing of indoor localization signals wherein statistical algorithms and machine learning are used in place of a fingerprint map. The disclosure relates to calculation of angle and distance based on measurements of an indoor localization signal, followed by energy-efficient distribution of signal processing. Local signal processing is performed using any of multiple eigen structure algorithms or a linear probabilistic inference, before cloud-based signal processing is performed using a nonlinear probabilistic inference and machine learning that's been trained with historical data transmitted by the base stations and time-of-day location patterns. Without having to generate and constantly update an energy-exorbitant fingerprint map, the disclosed system reduces localization error to merely 50 cm with 95% probability without compromising energy-efficiency to rival the accuracy of indoor localization systems that utilize fingerprinting.

The invention discloses an unmanned aerial vehicle radio interference suppression method and system based on causal reasoning, and relates to the technical field of radio direction finding and electromagnetic environment monitoring. The method comprises the following steps: carrying out probabilistic inference on interference causes under a structured generative causal model framework; and selecting a plurality of candidate communication actions in a preset intervention scheme library, evaluating the improvement effect on the link performance under each candidate communication action through anti-fact reasoning, selecting a target anti-interference action, and issuing and executing the target anti-interference action. According to the scheme, a systematic engineering solution from interference detection, cause diagnosis to strategy optimization and model updating is achieved, accurate identification, real-time anti-interference decision and long-term adaptive optimization of radio interference of the unmanned aerial vehicle in a complex electromagnetic environment are achieved, and the method has remarkable engineering application value and popularization significance.

A method and system for analysing and mitigating security risks in open innovation ecosystem is disclosed. The system comprises detects potential vulnerabilities in open innovation activities, including intellectual property exchanges, research data handling, and partner collaboration processes to provide identified threat data as input. The system represents interactions between an innovator and an adversary as a two-player zero-sum game. The system computes Nash equilibrium from the payoff matrix. The Nash equilibrium represents optimal defensive investments under adversarial conditions. The system also models adversary uncertainty using probability distributions, and further updates the equilibrium strategies based on incomplete or dynamic information. The system, thereafter, evaluates adversary uncertainty using Entropy-based risk assessment to determine levels of security investment resources responsive to the quantified uncertainty. Finally, the system integrates results of the equilibrium analysis, probabilistic inference, and uncertainty quantification to generate actionable security recommendations and guidelines for mitigation strategies.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using a quantum oracle to make inference in complex machine learning models that is capable of solving artificial intelligent problems. Input to the quantum oracle is derived from the training data and the model parameters, which maps at least part of the interactions of interconnected units of the model to the interactions of qubits in the quantum oracle. The output of the quantum oracle is used to determine values used to compute loss function values or loss function gradient values or both during a training process.

This application discloses a method and system for document review based on rule engine retrieval enhancement generation, relating to the field of document review. The method includes: initializing the review task's operating environment; performing structured processing and metadata binding on the document opening review rules corresponding to the target review task to construct a rule knowledge base; acquiring the target document text and performing preprocessing and updating; extracting factual information from the target document text and mapping the factual information to rule matching parameters; constructing a rule matching object based on the target document text, and using the rule matching parameters to call the rule engine retrieval enhancement generation mechanism to review the rule matching object and obtain the review explanation results; and summarizing and outputting a document opening review analysis report. This application avoids the rule omission problem caused by similarity retrieval, improves the reliability, consistency, and traceability of document compliance review, and reduces the fluctuations caused by probabilistic inference of the model.