933 results about "Uncertainty quantification" patented technology

The invention belongs to the technical field of bridge diagnosis, and discloses a bridge crack intelligent diagnosis system based on multi-modal data fusion. By fusing multi-source data such as visual images, sound wave detection and vibration signals, comprehensive perception and characterization of crack features are realized; constructing a bridge crack characteristic spectrum diagram by adopting a cross-modal feature extraction and heterogeneous feature coding technology; generating a crack evolution situation map based on space-time correlation analysis and knowledge graph construction; the robustness of the system in a complex environment is improved through environmental adaptability feature enhancement and multi-scale characterization; constructing a bridge safety risk hypergraph in combination with multi-dimensional risk analysis and multi-agent collaborative diagnosis; analyzing and revealing a crack evolution mechanism by applying a causal relationship; and finally, through dynamic fusion and uncertainty quantification, a crack intelligent diagnosis comprehensive report is generated. According to the system, the limitation of traditional single-mode diagnosis is broken through, and dynamic prediction and accurate risk assessment of fracture evolution are realized.

A federated distributed computational system enables secure oncological therapy optimization through robotic integration. The system establishes a distributed graph architecture with secure communication channels connecting computational nodes, implementing encryption protocols for cross-institutional data exchange. Each node contains processing capabilities for fluorescence-guided imaging, uncertainty quantification, and expert knowledge integration while maintaining hierarchical knowledge graphs of oncological biomarkers, interventions, and outcomes. The system coordinates domain-specific knowledge through token-space communication and implements an advanced robotic integration system for surgical interventions using spatiotemporal tumor mapping, multi-modal fluorescence imaging, surgical robot coordination, and space-time stabilized mesh management. Key capabilities include wavelength-specific multi-modal fluorescence detection, combined epistemic and aleatoric uncertainty estimation, tensor-based data integration with adaptive dimensionality control, and light cone search for adaptive treatment optimization—all while maintaining strict privacy controls.

The invention discloses a time sequence data modeling method for an agent with a body based on frequency domain learning, and the method comprises the steps: carrying out the self-adaptive partitioning of an original time sequence data sequence, extracting features through multi-scale convolution, and generating a position code; fourier transform is carried out on the feature block to extract frequency domain information, and nonlinear transform and residual compensation are carried out; calculating an attention relationship between the features and carrying out feature fusion; constructing a time sequence dependency relationship graph, and extracting time sequence aggregation features and dynamic system features; and constructing a probability prediction model, and calculating the uncertainty of a prediction result. According to the method, the modeling precision is improved through self-adaptive blocking and frequency domain feature extraction, the feature expression ability is enhanced through an attention mechanism and graph structure modeling, the prediction reliability is improved through uncertainty quantification, and the method is suitable for time series data modeling of an agent with a body in a complex environment.

The invention relates to the technical field of intelligent assessment and decision making, in particular to a robot dynamic risk assessment and decision making system and method based on multi-modal perception, and the system comprises a multi-modal sensor module which is used for collecting environment vision, acoustics, mechanics and position data in real time; the edge calculation unit is used for carrying out space-time alignment and feature fusion on the sensor data; the dynamic risk assessment model is used for integrating the environment uncertainty quantification module and the robot state prediction module based on a reinforcement learning framework; the decision execution interface is used for outputting a risk level and obstacle avoidance, speed reduction and shutdown instructions; by integrating visual, acoustic, mechanical and position multi-source sensor data and the like, the system can comprehensively capture various risk factors in a complex dynamic environment, so that the defect that a traditional single sensor system is insufficient in sensing dimension is overcome, and the system is particularly suitable for terrains and weather conditions with variable regions.

The invention discloses an industrial equipment fault prediction method based on multi-modal data, and belongs to the technical field of specific calculation models, and the method comprises the steps: carrying out the preprocessing according to the collected multi-modal data of the operation of industrial equipment, so as to unify the format of the multi-modal data, and obtaining the structural data; extracting features of the structured data one by one according to data categories, and obtaining a multi-modal fusion feature through a dynamic fusion mechanism; according to the multi-modal fusion features, a fault prediction classification score is obtained through a deep neural network model to perform fault prediction; and when the drift parameter of the multi-modal data is greater than a preset threshold value, performing incremental training on the deep neural network model through the multi-modal data collected in real time to update parameters of the deep neural network model. Through multi-modal data unified processing, dynamic feature fusion, deep neural network modeling and an online learning mechanism, the problems of insufficient multi-modal data fusion, prediction uncertainty quantization deficiency, poor model adaptability and the like are solved.

A deep learning system for time series prediction comprising a preprocessor that receives time series input sequences, truncates them by removing terminal values, and appends padding values to maintain the original sequence length. An encoder compresses these padded sequences into latent space representations, while a decoder reconstructs predicted sequences matching the original length, specifically trained to reconstruct values matching the removed terminal values in positions corresponding to the padding values. A training system optimizes the encoder and decoder by minimizing differences between original sequences and predicted sequences. The system can process multiple time horizons simultaneously while maintaining statistical properties and providing uncertainty quantification through confidence intervals. This approach enables accurate short-term forecasting while preserving both temporal patterns and statistical relationships in the predicted sequences.

The invention relates to the technical field of artificial intelligence and data processing, in particular to a mining explosion-proof high-voltage power distribution device-oriented all-insulation state intelligent monitoring method, which specifically comprises the following steps of: acquiring training data to construct a data set; marking the data in the data set; preprocessing the data in the data set; constructing an intelligent monitoring model for the full-insulation state of the mining explosion-proof high-voltage power distribution device based on a deep probabilistic neural network, inputting the preprocessed data set into the model for training, and repeating the iterative training process until a preset iteration stopping condition is met to obtain a trained model; and inputting the preprocessed real-time data into the trained model, generating probability distributions of normal operation, slight degradation and serious degradation and corresponding uncertainty quantized values by an output layer, and obtaining a final diagnosis result according to a preset discrimination condition. According to the invention, the accuracy and anti-interference capability of insulation state diagnosis are significantly improved.

The invention provides a sewage system traceability analysis and intelligent monitoring method, system and device based on a graph neural network, and a storage medium, and belongs to the technical field of environment monitoring and artificial intelligence. The invention aims to solve the technical problems of low efficiency, low precision, difficulty in processing multi-source data, poor monitoring network and the like of the existing sewage system pollution tracing method. The method comprises the following steps: constructing a sewage system knowledge graph fusing multi-source heterogeneous data such as water quality and water volume; adopting a multi-scale graph neural network model to learn pollution propagation characteristics based on the knowledge graph; after a pollution event occurs, pollution path backtracking is carried out in combination with physical models such as flow conservation so as to identify a pollution source; bayesian inference is introduced to carry out uncertainty quantification on a traceability result so as to assess the credibility of the traceability result; and finally, dynamically optimizing the layout of the monitoring points based on information gain and other criteria. According to the invention, rapid and accurate positioning of the pollution source can be realized, and the method is suitable for intelligent supervision of an urban sewage system.

The invention provides a thermal runaway prevention and heat dissipation optimization method and system based on a capacitor module, and relates to the technical field of capacitor module thermal management, and the method comprises the steps: extracting the multi-source operation data features of a single capacitor through an improved layered visual attention network, carrying out the data complementation through combining a bidirectional probability diffusion model, and obtaining a multi-source operation data feature of the single capacitor; and discovering a causal relationship between network identification data by using a causal structure, generating a performance analysis report, constructing a spatio-temporal dynamic graph neural network to predict temperature field distribution, performing fault diagnosis and uncertainty quantification in combination with a multi-task learning framework, further constructing a layered early warning mechanism, and performing early warning based on a layered reinforcement learning framework. And a dynamic prediction map and a performance analysis report are combined, a heat dissipation strategy is optimized, cooperative control of the heat dissipation units is realized, an optimal heat dissipation scheme is finally obtained, thermal runaway is effectively prevented, and the safety and reliability of the capacitor module are improved.

Processes and systems for automating a conveyance operation using an elongated conveyance member. In some embodiments, the process can include modeling one or more system state profiles of a downhole tool string during the conveyance operation within a wellbore to define one or more profile models; calibrating the one or more profile models; using the one or more calibrated profile models to calculate at least one of a tension profile and a force profile along the elongated conveyance member based, at least in part, on one or more detectable system states and / or one or more undetectable system states; and using one of: (i) a surface tension uncertainty quantification model to calculate a surface tension and uncertainty bounds around the calculated surface tension; or (ii) a surface weight uncertainty quantification model to calculate a surface weight and uncertainty bounds around the calculated surface weight.

The invention discloses a safe trajectory tracking method and system for a ground differential drive robot, and particularly relates to the technical field of robot operation management. Comprehensive means of multi-source data fusion, dynamic safe distance setting, trajectory prediction uncertainty quantification and trajectory deviation cost optimization are introduced; the path planning capability of the robot in a complex dynamic environment is improved; the behavior characteristics and environment complexity of the dynamic obstacle are evaluated in real time, the safe distance is dynamically adjusted, and the dynamic obstacle area is accurately quantified in combination with the trajectory prediction model, so that the problems of low trajectory tracking precision and insufficient obstacle avoidance robustness in the background technology are solved; the path optimization is realized through the minimization strategy of the trajectory deviation cost function, the operation safety of the robot in diversified scenes is ensured, and meanwhile, the efficiency and stability of path planning are greatly improved.

The invention provides a multi-station space-time correlation photovoltaic output probability prediction method and system based on interpretable deep learning, and relates to the field of photovoltaic power generation intelligent prediction. According to the method, a photovoltaic field station group output two-dimensional matrix is constructed based on photovoltaic output historical data, column vectors in the two-dimensional matrix serve as correlation test objects, and automatic typing is carried out on the correlation test objects. Fitting a probability distribution function corresponding to the photovoltaic output historical data after typing by using a non-parametric kernel density estimation method; and sorting out data containing space-time correlation information of the photovoltaic stations from the two-dimensional matrix, constructing a multi-station joint input matrix, and sending the multi-station joint input matrix into a bidirectional gating circulation unit for training to obtain an optimal photovoltaic output prediction model. And the model is used to complete photovoltaic output determinacy prediction in the future, and a probability prediction method is used to carry out uncertainty quantification on photovoltaic output. On the basis of data typing, kernel density estimation modeling is independently carried out for each type, the noise influence is reduced, and more accurate probability prediction is provided.

Systems and methods for determining an uncertainty measure associated with results of a medical task performed by an LLM (large language model) are provided. One or more prompts associated with a medical task are received. At least one of the one or more prompts are encoded into a set of features using a feature encoder network of an LLM. The medical task is performed based on the set of features using a decoder network of the LLM. An uncertainty measure associated with results of the medical task is determined based on the set of features using an uncertainty quantification module of the LLM. The results of the medical task and the uncertainty measure are output.

The invention relates to the field of heat energy engineering and automatic control, and discloses a multi-mode optimization system for combustion efficiency of a thermal power boiler. The system comprises a multi-modal data perception and space-time alignment module, a tensor manifold modeling and physical constraint feature extraction module, a space-time coupling dynamic prediction and uncertainty quantification module, a quantum optimization decision and DCS cooperative control module and a combustion state derivative early warning and optimization feedback module. Through multi-modal data space-time alignment, five-order tensor physical constraint modeling, PDE deep network prediction, quantum optimization decision and a closed-loop feedback mechanism, space-time unified fusion and physical feature extraction of combustion data are realized, the reliability of combustion state prediction is improved, an optimal control instruction is efficiently solved, system parameters are dynamically corrected, and the reliability of combustion state prediction is improved. The problems that in the prior art, data fusion is difficult, modeling physical constraints are lacked, optimization real-time performance is poor, and adaptivity is weak are solved, and the combustion efficiency and the intelligent control level of the thermal power boiler are remarkably improved.

The invention relates to the technical field of power grid operation and maintenance, and discloses an AI-based reactive compensation optimization method and system, and the method comprises the following steps: integrating the four-dimensional characteristics of traffic flow, power grid load, user preference and equipment health degree, constructing a space-time correlation graph model, achieving the adaptive evolution of a topological structure through dynamic graph convolution and an attention mechanism, and obtaining a reactive compensation optimization model. The method comprises the following steps: fusing spatial-temporal features of a graph neural network with a BiGRU time sequence for modeling, injecting an equipment health degree attenuation factor, generating a charging demand prediction result with uncertainty quantification, expanding an intelligent agent state space, improving a reward function, and embedding equipment health degree constraints in a strategy network and near-end optimization. By sensing the degradation state of the equipment in real time and intelligently adjusting the reactive compensation strategy, the service life of the old equipment is prolonged while the voltage stability is guaranteed, the deviation accumulation effect is effectively inhibited through feedforward-feedback composite control and dynamic rescheduling, and accurate matching of a scheduling instruction and actual output is ensured.

The invention relates to the technical field of intelligent traffic, and discloses a road transportation key operating vehicle intelligent supervision system, which comprises a data acquisition and fusion module used for acquiring multi-source heterogeneous data of a vehicle terminal, an external environment and the like in real time and carrying out standardization processing; the cognitive digital twinning construction module is used for constructing a dynamic heterogeneous graph representation human-vehicle-road-environment system and outputting a cognitive state vector through a graph neural network; the risk prediction and evaluation module is used for predicting a risk evolution trend based on the state vector sequence and quantifying the uncertainty of prediction by using a Monte Carlo discarding method; and the self-adaptive intervention decision module is used for combining risk prediction and uncertainty, making a decision through a reinforcement learning model and executing an optimal active intervention instruction. By constructing cognitive digital twinning and introducing uncertainty quantification and closed-loop self-checking, prospective prediction and self-adaptive intervention of driving risks are realized, and the accuracy and robustness of supervision are improved.

The invention relates to the field of water conservancy and hydropower engineering, in particular to a multi-objective optimized hydropower ecological scheduling decision-making system and method, and the system comprises a data collection module, an ecological model module, an intelligent decision-making engine module, a scheduling execution module, an effect evaluation module and a knowledge base module. The data acquisition module collects multi-source data, the ecological model module generates a training data set, the intelligent decision engine carries out ecological process modeling and probability prediction based on a deep learning architecture and spatial-temporal feature extraction, and generates a scheduling decision, the scheduling execution module controls hydropower engineering operation, and the effect evaluation module monitors ecological and economic effects. The knowledge base module stores historical experience and provides optimization suggestions, and the system improves the simulation accuracy of the ecological system, especially when the flow changes suddenly; and through uncertainty quantification, the system reliability and the ecological safety guarantee rate are enhanced, and high efficiency, accuracy and sustainability of ecological scheduling of the hydropower engineering are realized.

The invention relates to a multi-source data fusion-based urban physical examination evaluation index calculation method and system. The method comprises the steps of extracting a multi-source data sequence; identifying a data source of the urban physical examination index set, and extracting an independent time sequence data sequence; calculating the information entropy of the independent time sequence data sequence, and distributing a basic fusion weight; calculating a dynamic state evaluation value of the independent time sequence data sequence, and performing weighted fusion on the basic fusion weight and the dynamic state evaluation value to obtain a comprehensive state evaluation value; obtaining a distribution variance of the basic fusion weight, inputting the distribution variance into the uncertainty quantification model, and obtaining an index calculation result containing uncertainty measurement; the real-time performance of the evaluation result is enhanced through an aging attenuation mechanism, and the latest state of the city system is accurately reflected; the output uncertainty measurement index provides a quantitative basis of result credibility for a decision maker, and the decision risk caused by a data fusion error is reduced.

The invention relates to the technical field of coal mine telemetering, in particular to a mining laser methane telemetering system based on multispectral fusion, which comprises a multispectral laser emission module, an optical receiving and signal conversion module, a multispectral fusion processing module, an intelligent algorithm compensation module, a data transmission and display module and a power supply module. According to the method, accurate and real-time monitoring of the methane concentration in the coal mine environment is achieved, reliable guarantee is provided for coal mine safety production, compared with a traditional neural network compensation model, the optimized neural network compensation model has the advantages that the methane concentration compensation precision is greatly improved, the generalization ability of the model to different environment conditions is obviously enhanced, and the method is suitable for popularization and application. Accurate measurement of methane concentration can be realized in a more complex and changeable coal mine environment. Meanwhile, due to the application of a dynamic weight compensation mechanism and an uncertainty quantification and compensation adjustment method, the reliability and the stability of a measurement result are further improved, and a more powerful guarantee is provided for safe production of a coal mine.

The invention discloses a reservoir real-time scheduling simulation system based on a deep learning algorithm, and belongs to the technical field of intelligent water conservancy and artificial intelligence. Aiming at the problems of low prediction precision, poor multi-target coordination capability, weak coping uncertainty and the like of a traditional scheduling system, the system is designed to acquire hydrological, meteorological, water quality and engineering safety data through a multi-source data acquisition unit, and a multi-dimensional feature tensor is generated after preprocessing and fusion; the dispatching center server adopts an STGCN-LSTM mixed model to achieve high-precision prediction and uncertainty quantification of the water inflow process in the future 7-30 days, a reservoir hydrodynamic model and an MO-PPO algorithm are combined to complete multi-scene simulation and multi-target optimization decision, and an AF-DT mechanism dynamically adjusts the dispatching rule priority. According to the system, a sensing-decision-execution-feedback closed loop is constructed, the scheduling adaptive capacity and robustness are improved, the synergistic interaction of flood control, water supply, power generation and ecological protection is realized, and the system is suitable for real-time intelligent scheduling of large and medium reservoirs.

The invention discloses a string type photovoltaic inverter power prediction method and system, and relates to the technical field of photovoltaic inverters. According to the method, through dynamic abnormal data processing, multi-source time sequence alignment and multi-dimensional feature construction, the problems of large dimensional difference and time sequence asynchronization of original data are solved, and the generated fusion feature vector provides high-quality input for modeling and lays a prediction precision foundation; a dynamic response model is constructed in combination with a physical principle and data driving, targeted adjustment is achieved through a multi-working-condition feature library, and transient changes and non-ideal factor influences are accurately captured; the efficiency mapping model dynamic calibration mechanism further improves the power evaluation accuracy, solves the problem of poor working condition adaptability of a traditional model, constructs ultra-short-term, short-term and medium-term prediction sub-models, dynamically fuses the sub-models, combines error analysis and uncertainty quantification, outputs a multi-confidence prediction result, and improves the accuracy of power evaluation. Different time scale requirements of power grid dispatching, plan making and the like are met, and photovoltaic consumption and power grid stability are improved.

The invention belongs to the field of geotechnical engineering and engineering geological information modeling, and relates to a stratum-parameter coupling randomness three-dimensional random field modeling and shield construction ground surface settlement rapid prediction method and device based on multi-source data fusion. According to the method, a three-dimensional joint random field acting on stratum types and rock-soil key physical property parameters at the same time can be constructed under the joint constraint of multi-source data such as drilling, geophysical prospecting, geotechnical tests and terrains, and a calibrated, explainable and updatable geological section automatic generation and uncertainty quantification channel is formed; on the premise that geological rationality is guaranteed, section geometry and parameter distribution under the conditions of complex structures and lateral phase change are restored robustly, the accuracy, consistency and interpretability of the three-dimensional geological section are remarkably improved, and a combined uncertainty result which can be directly used for rapid prediction and risk assessment of shield construction settlement is output. The technical problems that a geological section generated by an existing method is not accurate enough, uncertainty is difficult to quantify, and the extrapolation capacity of a complex scene is weak are solved.

The invention discloses a digital twin-driven bridge full life cycle damage prediction and evaluation method and system, and belongs to the field of bridge structure health monitoring, and the method comprises the steps: obtaining a monitoring data set of a bridge structure; an initial digital twinborn model embedded with a micro physical layer is constructed and trained, the micro physical layer constructs a damage evolution model applied with monotonic physical constraint based on multi-source monitoring data, and a damage evolution trajectory of the bridge in a future time period is predicted through the damage evolution model based on the physical parameter vector; in combination with an uncertainty quantification method, generating a time-varying reliability index of the bridge in a future time period; and based on the time-varying reliability index, constructing and solving a maintenance decision optimization model to generate a maintenance decision of the bridge. According to the invention, the physical authenticity and reliability of the long-term prediction result are ensured.

The invention relates to the technical field of machine vision, and discloses a surveying and mapping geographic information analysis method and system based on machine vision and a medium. The method comprises the following steps: preprocessing multi-source remote sensing data to obtain a standardized surveying and mapping data set, and extracting ground feature edge and texture information to obtain a multi-scale spatial feature and a semantic segmentation result; carrying out multi-source data integration to obtain a unified geographic element representation diagram, and carrying out geographic element identification on the unified geographic element representation diagram to obtain a surveying and mapping geographic element set; performing Bayesian inference of spatial uncertainty quantization and parameter uncertainty quantization on the surveying and mapping geographic element set to obtain a comprehensive uncertainty evaluation result; and constructing vectorized geographic information based on the surveying and mapping geographic element set and the comprehensive uncertainty evaluation result. According to the method and the device, the edge information of the geographic elements is effectively reserved, the identification precision of the geographic elements is improved, the geographic information expression is more structured, and automatic updating is realized.

The invention relates to the technical field of data processing, and discloses an ocean current simulation method and system based on a three-dimensional temperature-salinity value. The method comprises the following steps: carrying out line calculation and adaptive mesh refinement processing on a three-dimensional temperature and salt field simulation model to obtain a model operation scheme; performing three-dimensional visualization processing on the three-dimensional temperature and salt field initial data set and the model operation scheme to obtain dynamic three-dimensional temperature and salt field visualization data, and analyzing the dynamic three-dimensional temperature and salt field visualization data to obtain three-dimensional structure data and evolution process data of the key ocean phenomenon; performing data assimilation processing on the multi-source real-time observation data, the three-dimensional structure data and the evolution process data to obtain a fusion data set; and performing ocean current prediction calculation on the fused data set to obtain a three-dimensional ocean current prediction field, and performing uncertainty quantitative analysis on the three-dimensional ocean current prediction field to obtain ocean current simulation probability distribution data. According to the invention, the accuracy and timeliness of ocean current simulation based on the three-dimensional temperature-salinity value are improved.

The invention relates to the technical field of smart factory decision-making, and discloses a digital twin-driven smart factory AI intelligent decision-making system, which comprises a data acquisition and preprocessing module used for acquiring and preprocessing real-time and historical data of a production line; the twinborn modeling identification module is used for constructing a twinborn body and carrying out parameter identification and uncertainty quantification; the alignment evaluation credible module is used for comparing twin prediction data with real data and generating availability marks and trust scores; the strategy simulation optimization module is used for generating candidate strategies and risk evidences based on the constraints and the key performance indicators; the grayscale online publishing module is used for screening and grayscale publishing a strategy according to the trust score and the performance result; the operation evaluation auditing module is used for collecting operation data and generating an auditing packet; and the drifting detection updating module is used for detecting distribution drifting and recalibrating and updating the twinborn body and the strategy library. According to the invention, the reliability, robustness and continuous optimization of the production decision of the smart factory are realized.

The invention discloses a metal casting process quality monitoring system for intelligent manufacturing and a method thereof, and relates to the technical field of metal casting quality monitoring. According to the system, a Bayesian method and multi-model hypothesis testing are used for monitoring the casting process in real time, whether deviation or abnormity exists or not is recognized, and corresponding adjustment is made, so that it is ensured that the quality of a casting always meets the preset standard; performing parallel processing through four pre-training models, performing feature extraction and fusion on the sensor data after standardization processing, and performing fine adjustment on the pre-training models through transfer learning; and performing classification prediction or regression prediction on the fused feature vectors based on an integration method of a multi-layer perceptron model, and evaluating the uncertainty of a prediction result in combination with an uncertainty quantification method. By means of the system and method, various physical parameters in the casting process can be monitored in real time, accurate prediction can be achieved, problems can be found in time, waste in production is reduced, and the overall quality of castings is improved.

The invention discloses a radar lifting control method and system based on meteorological monitoring, and relates to the technical field of radar lifting control, and the method comprises the steps: completing the switching of a power supply and communication after a radar is powered on, initializing a controller, collecting the data of a meteorological station, and generating a future fusion wind speed in real time through a Kalman filtering physical model and a residual neural network; future fused wind speed is converted into wind pressure for evaluation, the risk degree is judged according to the evaluation result, early warning is given out, and the controller is preheated to enter a lifting preparation state. The input stability is improved through meteorological data sliding window smoothing and feature extraction, wind speed dynamic prediction and uncertainty quantification are achieved through XGBoost prediction and residual variance estimation, the time sequence consistency and robustness are enhanced through remote API interpolation correction and adaptive extended Kalman filtering, residual correction is conducted through a neural network, the prediction precision is improved, and the prediction accuracy is improved. And a reliable decision basis is provided for radar lifting control.

The invention discloses a railway infrastructure health state prediction method, device and platform. The method comprises the following steps: collecting multi-mode sensor data reflecting the state of the railway infrastructure; the data of the multi-mode sensor are preprocessed; constructing and training a multi-modal data fusion deep learning model; and inputting preprocessed data collected in real time into the trained multi-modal data fusion deep learning model to obtain a health state prediction result of the railway infrastructure, and comparing the health state prediction result with a preset health threshold. According to the technical scheme, data standardization is achieved by collecting the data of the multi-modal sensor and adopting the preprocessing technology; a multi-task deep neural network self-organizing mapping and a graph neural network are used for fusing modal features, and a multi-task long-short-term memory network and a Bayesian uncertainty quantification module are used for assisting, so that the railway health state is accurately predicted.

The invention relates to the technical field of medical image processing, in particular to a CT image intelligent analysis system for pneumonia auxiliary screening. The method comprises the following steps: firstly, preprocessing a chest CT image and detecting a candidate focus area; secondly, extracting a topological feature, a deep convolution feature and a texture statistical feature based on a persistent coherence theory from each candidate focus, and performing feature fusion through a multi-head self-attention mechanism to generate a unified focus representation vector; mapping the lesion characterization vectors to a pre-constructed radiology knowledge graph, adopting a graph neural network for reasoning, and outputting the pneumonia suspected probability and lesion classification of each lesion; and finally, performing fusion and uncertainty quantification on the analysis results of the plurality of focuses by adopting an evidence theory, and generating a comprehensive screening report. According to the method, complex-form lesions are effectively identified through topological features, accurate identification of lesion types is realized through knowledge graph reasoning, and diagnosis uncertainty quantification is provided through an evidence theory.