23results about How to "Fully portrayed" patented technology

The application discloses a kind of based on big language model and ability atlas's wisdom pastor authentication examination method and system, belong to artificial intelligence technical field.The application first constructs the ability atlas based on knowledge, skill, behavior three-dimensional ontology model;Obtain examinee multimodal data and generate initial ability vector;Through double-path dynamic gate graph neural network, real-time update ability vector, the network includes erase gate and increase gate and is updated by graph convolution layer propagation;According to ability short board, determine aggressive factor and combine incidental factor, use big language model to dynamically generate the next logical unique test question;Repeat interaction until the end of examination;Finally, according to all updated ability vector, generate digital ability portrait report.The application realizes dynamic individualized surveying and mapping, anti-cheating and deep characterization of talent ability structure, improves the precision and security of wisdom pastor authentication examination.

The application discloses an emotion recognition method based on online cross-modal knowledge distillation, comprising the following steps: acquiring electroencephalogram and electrocardiogram original signals and windowing and cutting; constructing electroencephalogram and electrocardiogram student models, extracting intermediate features from each modal data segment through an encoder, and obtaining non-normalized prediction output through a classifier; constructing a teacher probability distribution through a joint encoder fusion; introducing adaptive contrast loss to align the cross-modal intermediate features, introducing distillation loss to constrain the prediction probability distribution of each modal to align with the teacher probability distribution; synchronously optimizing new student model parameters through online collaborative training; and performing actual inference prediction based on the student model after training. The application combines double modal signals to make up for the defects of single modal information, excavates the complementarity of modes, realizes dynamic generation of teacher supervision signals and real-time collaborative learning of modes through online distillation, does not increase test calculation overhead, effectively improves the recognition accuracy, model robustness and generalization ability, and has good application prospect.

The invention discloses an underwater AUV (Autonomous Underwater Vehicle) service caching and switch state switching method based on a D3QN (Digital 3QN). The method comprises the following steps: firstly, for an ocean network formed by a water surface buoy node and an AUV, establishing an AUV service cache state and dormancy / activation state coupling control model; then, through consistent binding of a target switch state and a service cache write-in action, dormancy isolation constraint and a dormancy cache freezing rule, a feasible action mask meeting cache capacity constraint and anti-repeated write-in constraint is constructed; the method comprises the following steps of: firstly, constructing a comprehensive cost function according to cache update time overhead, write-in energy consumption, switch state switching cost and invalid activation penalty, and finally, solving in a legal action subspace by using D3QN, and outputting an AUV service cache and switch state switching control strategy. According to the invention, AUV service cache updating and switch state switching under the dynamic load can be effectively realized, and underwater network energy consumption and service cache updating delay are reduced.

The invention relates to an online community creator feedback prediction method and system based on a dynamic reputation graph and text analysis. Relates to the technical field of feedback prediction. The method comprises the following steps: S1, acquiring and preprocessing historical interaction data containing comment texts, user IDs, timestamps and like and treading records; s2, constructing a like and treading double-view interaction map; s3, segmenting the data according to the day, and calculating a daily positive and negative reputation value through a PageRank algorithm; s4, introducing a time decay function, and carrying out weighted summation on the daily granularity reputation value in the close time window to obtain a dynamic reputation; s5, text semantic features are extracted through RoBERTa, the text semantic features and the dynamic reputation features are spliced and fused, and the full-connection neural network is input to output the like or treading probability. The method gives consideration to positive and negative reputation and time dynamics, makes up for the defects of plain text prediction, remarkably improves the feedback prediction precision, especially optimizes the click behavior prediction effect, and can provide support for community atmosphere guidance and network violent early warning.

The invention relates to a new energy output prediction method and system based on an NHITS model, and the method comprises the steps: obtaining the input characteristics of a to-be-tested new energy object, forming new energy time series data, and enabling the new energy to comprise photovoltaic or wind power; the input characteristics corresponding to photovoltaic output prediction comprise time, solar irradiance, air temperature, air pressure, humidity and historical power data, and the input characteristics corresponding to wind power output prediction comprise time, meteorological variables and wind speeds and wind directions at different heights; new energy output prediction is carried out by adopting an NHITS prediction model, the NHITS prediction model carries out multi-time scale decomposition on a time sequence through a hierarchical recursive structure, the time sequence is divided into a plurality of modules stacked according to layers, and each module carries out partial interpretation on input new energy time sequence data under different time scales and outputs a prediction component; and superposing the prediction components of each layer to obtain a final prediction result. Compared with the prior art, the method can maintain high prediction precision and stability.

This invention relates to the field of fault early warning technology, and provides a method and system for early warning of faults in large oil-filled equipment, comprising: Step 1, collecting acoustic fingerprint data throughout the entire lifecycle of the large oil-filled equipment and constructing an acoustic fingerprint sample set; Step 2, preprocessing the acoustic fingerprint sample set, extracting acoustic fingerprint feature parameters, and establishing a standardized acoustic fingerprint database; Step 3, constructing a multi-channel deep learning acoustic fingerprint recognition model and a discharge severity assessment model based on deep learning; Step 4, based on the output of the trained acoustic fingerprint recognition model and combined with a preset multi-level early warning threshold system, performing real-time assessment of the operating status of the large oil-filled equipment, and generating corresponding early warning information when the assessment result meets the early warning triggering conditions; Step 5, pushing the early warning information to the operation and maintenance terminal, and automatically generating operation and maintenance suggestions based on the defect type and severity. This invention can effectively provide early warning of faults in large oil-filled equipment.

The invention discloses an emotion recognition method based on online cross-modal knowledge distillation. The emotion recognition method comprises the steps that electroencephalogram and electrocardio original signals are obtained and subjected to window segmentation; constructing an electroencephalogram and electrocardio student model, extracting intermediate features from each modal data segment through an encoder, and obtaining non-normalized prediction output through a classifier; teacher probability distribution is constructed through joint encoder fusion; self-adaptive comparison loss alignment cross-modal intermediate features are introduced, and distillation loss is introduced to constrain prediction probability distribution of each modal to align to teacher probability distribution; new student model parameters are optimized synchronously through online cooperative training; and performing actual reasoning prediction based on a student model after training. According to the method, bimodal signals are combined to make up single-modal information defects, modal complementarity is mined, teacher supervision signal dynamic generation and modal real-time collaborative learning are realized through online distillation, test calculation overhead is not increased, identification precision, model robustness and generalization ability are effectively improved, and the application prospect is good.

The invention relates to the technical field of neural image analysis, and particularly provides a brain function connection intelligent screening method and system for autism spectrum disorders, and the method comprises the steps: firstly obtaining resting state functional magnetic resonance imaging data and phenotype information thereof, and extracting a blood oxygen level dependence value time sequence of each brain region after preprocessing; then constructing a multi-scale brain network comprising a low-order function connection matrix and at least one high-order function connection matrix; the matrix is converted into a brain function connection graph containing sub-graphs of different scales through threshold sparsification; meanwhile, phenotype embedding features are extracted from phenotype information; the graph data and the phenotypic features are input into a multi-channel neural network model for parallel processing and fusion, and joint feature representation is obtained; and finally, outputting an auxiliary diagnosis result of the autism spectrum disorder through the classifier. According to the method, by fusing the multi-scale brain function connection information and the individual phenotype features, the accuracy of autism classification diagnosis and the generalization ability of the model are effectively improved.

The invention belongs to the technical field of power distribution network operation fault diagnosis, and particularly relates to a district voltage quality digital portrait construction and low-voltage fault identification method and system.The district voltage quality digital portrait construction comprises the steps that firstly, a low-voltage fault evaluation index system is constructed, and index data are collected; then, performing multi-scale time-frequency analysis on the acquired voltage signals by adopting different signal processing algorithms, splicing and fusing output results of the different signal processing algorithms to form a fused frequency domain feature vector, and jointly using the fused frequency domain feature vector and each index data as a digital portrait for describing the voltage operation state of the transformer area; according to the transformer area voltage low-voltage fault identification method, a low-voltage fault identification model is firstly constructed, and the model realizes low-voltage fault identification based on a digital portrait. According to the method, the operation characteristics of the power distribution area can be comprehensively and visually described through the digital portraits, high-quality input is provided for follow-up fault recognition, and finally the fault recognition precision is improved.

The application provides a well-ground differential electromagnetic three-dimensional inversion method considering the effect of well casing, and relates to the geophysical logging technical field; an additional term is added to modify the average resistivity formula, so that numerical simulation containing the steel casing is realized without refining the grid, and the resistivity model containing the casing effect is inverted by the Gauss-Newton method. The resistivity model parameters containing the casing effect are inverted, the well-ground differential three-dimensional electromagnetic inversion considering the effect of well casing is realized, and the precision and accuracy of the well-ground electromagnetic inversion are further improved.

This invention relates to the field of intelligent lighting technology and discloses a method and system for predicting urban lighting demand based on a spatiotemporal graph neural network. The system includes: a multi-source data perception and processing module, a spatiotemporal graph construction and feature learning module, a meta-learning adaptive prediction module, a lighting strategy optimization and execution module, and a model online update module. By integrating multi-source heterogeneous data from IoT terminals and urban management databases to construct a unified spatiotemporal dataset, and utilizing a spatiotemporal graph neural network to simultaneously learn the complex dependencies of different nodes in the city in the temporal and spatial dimensions, it can more comprehensively and accurately depict the formation and evolution of urban lighting demand, thereby improving the accuracy and reliability of predicting future lighting demand intensity and providing a basis for energy-saving regulation. By introducing a meta-learning strategy, the prediction model can utilize historical experience to adapt to new urban areas and time periods, solving the problem of insufficient generalization ability of the model in unfamiliar scenarios.

The invention discloses a passenger transport airport service range prediction method based on multi-source data fusion, and relates to the technical field of airport service information processing, and the method comprises the steps: constructing a starting point individual attribute feature vector, an airport terminal point attribute feature vector, and constructing a spatial impedance variable; taking the starting point individual attribute features, the ending point airport attribute features and the spatial impedance variable as input, constructing and training a depth gravity model, applying the trained depth gravity model to global space grid units of a research area, calculating the probability of each grid unit for selecting each airport, and calculating the airport selection probability of each grid unit; determining an affiliation airport of each grid unit according to a probability maximum principle so as to generate continuous space service range distribution; the problem that a traditional airport service range prediction method is low in prediction precision in a multi-airport competition environment is solved.

This invention discloses a rapid detection method and system for the reheating degree of meat dishes based on electrochemical sensing, belonging to the field of food testing technology. Addressing the problem of difficulty in rapidly and accurately evaluating the reheating degree of meat dishes in existing technologies, this invention first places the meat dish to be tested in a space, generating headspace gas at a preset equilibrium temperature and time. An electrochemical sensor unit collects the complete time-response curve of the interaction between the headspace gas and the sensitive layer, extracting dynamic characteristic parameters such as response rise time, peak response value, time to reach peak value, curve integral area, and half-life of the fall phase. Simultaneously, it acquires ingredient attribute parameters and reheating method parameters. These parameters are then input into a pre-trained deep learning prediction model, which outputs the reheating degree level and displays it on a display unit. This method can be used in the catering industry, food processing, and home cooking for rapid detection of the reheating degree of meat dishes, ensuring food quality and food safety.

The application discloses an AI image discrimination method based on frequency domain and noise domain consistency, and belongs to the technical field of image processing, computer vision and artificial intelligence security. The method comprises the following steps: acquiring an image to be discriminated and performing pretreatment, converting the image into a brightness channel and acquiring frequency spectrum information; inputting the image into a dual-domain physical guidance feature extraction module to extract directional spectrum features and multi-scale noise flow features; inputting the directional spectrum features and the multi-scale noise flow features into a cross-domain physical consistency module to generate a local consistency distance map and a global consistency score through shared embedding mapping and feature distance calculation, and obtaining consistency perception features; inputting the directional spectrum features, the multi-scale noise flow features and the consistency perception features into a dual-flow fusion module for cross-flow interactive fusion and global feature aggregation to obtain discrimination feature representation; and inputting the discrimination feature representation into a classification head to output a discrimination result of whether the image to be discriminated is a real image or an AI generated image. The application realizes effective discrimination of AI generated images by jointly modeling the physical consistency relationship between frequency domain features and noise domain features, and improves the robustness and generalization ability of the model under the conditions of cross-generator, cross-dataset and image compression, blurring and other post-processing conditions while ensuring detection accuracy.

The application belongs to the technical field of traffic flow prediction, and particularly relates to a spatio-temporal traffic flow prediction method fusing multi-graph structure and knowledge enhancement, comprising the following steps: obtaining historical observation data of traffic monitoring nodes at continuous multiple time steps; constructing enhanced features based on the historical observation data; splicing the enhanced features and the historical observation data to obtain enhanced input features; inputting the enhanced input features into a pre-trained traffic flow prediction model; and the traffic flow prediction model predicting traffic flow at one or more future time steps. The application realizes joint modeling of multi-source spatial correlation, complex time dynamics and historical period knowledge in traffic flow data, and improves the accuracy and robustness of future multiple time step traffic flow prediction.

The invention relates to the technical field of electroencephalogram signal monitoring, in particular to an electroencephalogram signal abnormal state recognition method based on theta-gamma cross frequency coupling characteristics, which comprises the following steps: acquiring an intra-operative electroencephalogram signal of a patient and segmenting the electroencephalogram signal to obtain segmented electroencephalogram signals; filtering, denoising and self-adaptive artifact elimination processing are carried out to obtain a preprocessed electroencephalogram signal; extracting multi-dimensional electroencephalogram features from the preprocessed electroencephalogram signals, extracting alpha-wave average power, theta-wave phase, normalized instantaneous frequency deviation, gamma-wave amplitude envelope and theta-wave average power by adopting Hilbert transform, calculating theta-gamma coupling strength, constructing a multi-dimensional electroencephalogram feature vector, and calculating the multi-dimensional electroencephalogram feature vector; meanwhile, the explosion suppression ratio of the preprocessed electroencephalogram signals is calculated to serve as an auxiliary feature; and performing anomaly identification based on the multi-dimensional anesthesia state feature vector and the auxiliary feature. According to the technical scheme, the misjudgment risk caused by single characteristic fluctuation can be reduced, and the reliability of electroencephalogram signal abnormity monitoring in the anesthesia process is improved.

This invention discloses a digital health diagnosis method for diesel engines based on multi-source data fusion, comprising the following steps: collecting signals generated during diesel engine operation and preprocessing them to generate multi-source combustion response data sequences; mapping events to combustion propagation nodes to generate a set of combustion propagation nodes; establishing propagation connections between nodes to generate a set of combustion propagation paths; constructing a combustion propagation topology; calculating the frequency of combustion propagation path occurrences and the stability of propagation time differences to generate a stable domain for the combustion propagation structure; performing structural matching between the current combustion propagation topology and the stable domain for the combustion propagation structure to generate a topology deviation structure; determining the diesel engine health status based on the topology deviation structure and generating a digital health diagnosis result for the diesel engine. This invention utilizes multi-source data fusion and combustion propagation topology analysis methods to achieve abnormal identification of diesel engine combustion structures, possessing advantages such as high diagnostic accuracy and strong structural characterization capabilities.

The invention relates to a TMD detection model based on voice time-frequency feature fusion, and a construction method and system thereof, and the model comprises a time sequence feature extraction module which is used for extracting the time sequence features of a voice signal according to MFCC features; the frequency domain global feature embedding module is used for embedding the acoustic features related to the TMD in the frequency domain of the voice signal into the time sequence features; the classification module is used for judging whether the patient corresponding to the voice signal is a TMD patient or not; speech signals are extracted, MFCC features and acoustic features of the speech signals of a TMD patient and a non-TMD patient are input into a time sequence feature extraction module and a frequency domain global feature embedding module respectively, an initial TMD detection model is trained, and a TMD detection model used for detection is obtained; according to the method, the overall feature expression capability is enhanced, and the accuracy and robustness of the TMD detection model are remarkably improved, so that the recognition effect of the TMD detection model on TMD-related voice anomalies is improved, and pathological information hidden in voice signals can be more comprehensively mined.

The invention discloses a recommendation object optimization method and device, electronic equipment and a program product, and relates to the technical field of content recommendation, the method comprises the following steps: obtaining theme features of a recommendation set, the recommendation set being used for representing multi-modal recommendation information; obtaining a matching relationship between the theme feature and an interest portrait of the first recommendation object, wherein the interest portrait is determined based on first interaction data of the first recommendation object; screening the first recommendation object based on the matching relationship to obtain a second recommendation object; recommending the recommendation set to a second recommendation object, and obtaining second interaction data of the second recommendation object for the recommendation set; and optimizing the second recommendation object by using the second interaction data. By implementing the technical scheme of the application, the target users can be screened and recommended based on accurate matching of the recommended collection theme and the user interests, and then the target user group is continuously and dynamically optimized by using the feedback data, so that the recommendation accuracy and the interaction conversion rate are effectively improved.

The application discloses a domain-invariant multi-modal emotion recognition method based on multi-time scale contrast learning, which comprises a pre-training stage and a fine-tuning stage; wherein the pre-training stage is to input the physiological signals of M modalities of different subjects into a modal encoder to extract features, and to align the multi-modal physiological signals under the same emotional stimulus through time contrast learning and cross-modal contrast learning to obtain emotion-related representations; meanwhile, the high-dimensional features output by the encoder are learned through an adversarial domain adaptation mechanism to suppress the information related to the subject identity in the features and to reduce the influence of individual differences, so as to obtain multi-modal feature representations which are invariant to subjects. The fine-tuning stage is to model the multi-scale time features by using the encoder obtained through pre-training, and to complete emotion recognition training through modal fusion and a classifier. The application can effectively improve the emotion recognition capability of the model in the cross-subject scenario.

This invention relates to the field of image data processing technology, and more particularly to a method and system for visually monitoring the operating status of an energy-feedback elevator. The method includes: acquiring operating images of key components in the energy-feedback elevator; determining the local structural orientation angle of each pixel; determining the structural anomaly value of each pixel to obtain a structural anomaly feature map; constructing a weighted gray-level co-occurrence matrix in all directions at a preset distance, and determining the frequency of all gray-level pairs in the weighted gray-level co-occurrence matrix; determining whether there are surface defects in the key components of the energy-feedback elevator, classifying and evaluating the surface defects of the key components, and visually displaying the evaluation results on a monitoring interface. This invention analyzes the structural orientation angle based on pixel gradients, and combines the gradient magnitude with the neighborhood angle to evaluate anomalies, effectively distinguishing between normal textures and defects, and improving the accuracy of monitoring defects in periodic texture components.

This invention provides a method for analyzing the spatial transmission effect of flood control resilience in watersheds, involving the interdisciplinary fields of watershed flood control analysis, spatial econometric analysis, and water resources management. The method includes: obtaining a set of indicators to be used based on multi-source basic spatial data; acquiring the flood control resilience influencing factor vector and flood control resilience value for each administrative unit; correlating and mapping the flood control resilience influencing factor vector and flood control resilience value at the administrative unit scale to obtain the flood control resilience influencing factor vector and flood control resilience characterization value for each sub-watershed; constructing a spatial weight matrix and a spatial econometric model to identify target sub-watersheds with spatial transmission effects; and characterizing the impact of spatial transmission effects at the administrative unit scale based on the model parameter estimation results for the target sub-watersheds and all sub-watersheds. This invention can clearly identify the main driving factors of flood control resilience within a watershed and their spatial transmission paths.

A method for predicting coal spontaneous combustion temperature based on adaptive fusion of multi-source heterogeneous data is proposed. This method collects four data sources: time-series data of indicator gas concentrations, distributed fiber optic temperature field data, infrared thermal imaging data, and environmental parameter data. Modal features are extracted using bidirectional long short-term memory networks, one-dimensional convolutional networks, two-dimensional convolutional networks, and fully connected networks, respectively. Data quality scores are calculated in real-time for each data source, generating adaptive fusion weights. A cross-modal multi-head attention mechanism is used to deeply fuse the multi-modal features. The fused features are then propagated multiple times forward using a Monte Carlo random deactivation method, outputting the predicted temperature value and its confidence interval. A risk assessment index is constructed by combining the temperature rise rate, enabling graded early warning of coal spontaneous combustion temperature. This invention overcomes the shortcomings of existing methods, such as reliance on a single data source, lack of sensor fault tolerance, and lack of confidence assessment for prediction results, thus improving the accuracy, robustness, and scientific validity of coal spontaneous combustion temperature prediction and early warning.