15 results about "Feature model" patented technology

The application discloses a kind of civil aviation safety report theme modeling methods based on large language model and semantic enhancement, its method includes: civil aviation safety theme semantic feature model extracts the text semantic vector and event structure vector of safety report text, and obtains the fusion feature vector of safety report text by fusion;Safety report text database is carried out cluster clustering processing and obtains initial cluster set and noise sample set;Select representative front p% as the representative sample set of cluster;Construct noise sample evaluation repair mechanism module, and select the candidate cluster to which noise sample belongs using noise sample evaluation repair mechanism module;Comprehensive gain function is constructed using representative sample set in cluster, and representative sample iteration screening processing of candidate sample is carried out in cluster.The application realizes the theme modeling goal of semantic accuracy, comprehensive coverage and stable result by multi-module collaborative innovation, and provides reliable technical support for management decision.

The method includes receiving spectral data of a substrate and measurement data corresponding to the spectral data of the substrate. The method includes determining multiple feature model configurations for each of a plurality of feature models, further including determining that each of the multiple feature model configurations includes one or more feature model conditions. The method includes determining multiple feature model combinations, further including determining that each of the multiple feature model combinations includes a subset of the multiple feature model configurations. The method includes generating multiple input datasets, each of which is generated based on applying spectral data to each of the multiple feature model combinations. The method includes training multiple machine learning models, each of which is trained to produce an output using an input dataset from a plurality of input datasets and measurement data. The method further includes selecting from the trained machine learning models a trained machine learning model that satisfies one or more selection criteria.

The application relates to the technical field of computers and discloses a full-movement simulator fault self-recovery optimization method and system, an electronic device and a storage medium. The method comprises the following steps: collecting multi-dimensional state data, identifying a fault type and locating a root cause, performing cross verification and confirmation through remote display picture and local rendering data comparison, performing hierarchical self-recovery according to the fault type and the severity, backing up configuration, cache and log to support rollback, performing bidirectional connectivity test and link health degree evaluation on network faults, executing adapter reset, driver reload and master / standby link / network port switching and outputting deterioration early warning, grading and caching terrain data according to distance, preloading tiles in combination with flight situation to predict a path range, regularly cleaning the cache and synchronizing with a data source, controlling a start timing based on a node dependency relationship and adjusting parameters in a maintenance mode, training a time sequence feature model based on historical fault data to realize trend early warning and preventive maintenance, and archiving disposal data to update a diagnosis rule and a self-recovery strategy.

This invention presents an AI-generated text detection method based on graph structure features, belonging to the fields of artificial intelligence and natural language processing. The method includes: dataset construction, entity relation extraction and graph structure construction, graph structure feature extraction, graph structure feature model training, and text detection. It further incorporates traditional text feature extraction and model training, adaptively fusing the traditional text feature model and the graph feature model based on confidence-weighted entropy, and then performing text detection based on the fused model. This invention is the first to perform AI text detection from the perspective of graph structure features, breaking through the limitations of existing research that focuses on surface features such as vocabulary, syntax, and perplexity. The fusion strategy dynamically adjusts the fusion weights by quantifying the uncertainty of model predictions, maintaining a high level of performance on both original data and adversarial examples, achieving a balance between detection accuracy and adversarial robustness. It can be widely applied to the detection of AI-generated content such as news content and academic papers.

The application provides a delay feature extraction and multiplexing method, device, equipment and readable storage medium, the method comprises the following steps: for each original delay sequence, the original delay sequence is divided into a plurality of original subsequences, and a plurality of groups of original description parameters are determined; for each original subsequence, the original description parameters and the source information are structured and packaged to obtain a feature model; according to the IP level test requirement, one or more feature models of a target IP core are taken as target models, for each target model, the target description parameters are determined according to the IP level test requirement and the original description parameters, the target subsequence is generated according to the target description parameters; when there are a plurality of target models, a plurality of target subsequences are concatenated in the order of the corresponding plurality of original subsequences to obtain a target delay sequence. Through the application, the delay data required by different IP level tests can be accurately obtained without repeatedly performing SoC level tests.

This invention provides a distributed industrial process data imputation method based on lightweight heterogeneous graph neural networks. Based on real industrial process time-series data, this invention establishes a lightweight heterogeneous graph neural network distributed data imputation method. It utilizes a lightweight multi-layer linear model to extract rich deep features from the time series data and leverages various high-order heterogeneous information to learn spatial relationships, thereby imputing missing values. This lightweight heterogeneous graph neural network data imputation method can solve the problems of distributed industrial process data, complex high-order heterogeneous information among the data, which makes data feature modeling difficult and spatial information capture insufficient. It achieves high-precision distributed industrial data imputation, providing technical support for industrial engineering.

This invention relates to the field of workpiece recognition technology, and particularly to a method and system for intelligent workpiece recognition based on graphic description. The method includes: reading a 3D drawing and extracting geometric element information and product manufacturing information; determining the adjacency relationships between surfaces and calculating their included angles based on the extracted geometric element information, and identifying independent processing features on the workpiece; matching the identified independent processing features with a pre-built feature model library to determine the process template corresponding to each independent processing feature and calculating the corresponding processing dimensions; calculating the processing time for each independent processing feature based on the processing dimensions and a preset process parameter library, summing the total processing time, and calculating the total processing cost of the workpiece. This invention matches corresponding processing process templates to different processing features, thereby achieving automatic calculation of processing costs. It enables cost calculation before workpiece production, improving the accuracy of cost control.

This invention relates to the fields of image processing and computer vision, specifically to a hierarchical multi-scale semantic segmentation convolutional neural network, comprising an encoder, a decoder, a connection module, a feature extraction module, a feature enhancement module, and a feature fusion module. The encoder extracts features from the input image and outputs multi-layer feature maps. The decoder upsamples layer by layer to restore feature resolution. The connection module fuses features from different levels. The feature extraction module performs multi-scale processing on the input features through multiple parallel convolutional branches. The feature enhancement module performs global feature modeling on the fused features. The feature fusion module processes and fuses the features using a parallel structure of symmetric and asymmetric convolutional kernels. This invention enables automatic building extraction in high-resolution remote sensing images, achieving both high accuracy and high efficiency, and solves the problem of robust segmentation under complex scenes, multi-scale targets, and resource-constrained conditions.

The application discloses a kind of based on deep learning's reflective metasurface phase prediction method.The method includes: using 16×16 discrete coding symmetric structure to construct metasurface unit, improve the design space of metasurface, and generate the data set including structure image and corresponding reflection phase real part and imaginary part curve by Python-CST joint simulation;A kind of hybrid neural network model is constructed by local feature extraction module, global feature modeling module and regression output module, wherein local feature extraction is realized based on ResNet, and global feature modeling is completed by introducing position coding and frequency coding Transformer encoder;While designing the complex loss function including phase real part error, imaginary part error and physical constraint term, and the weight coefficient thereof is systematically optimized to balance prediction accuracy and physical rationality;Optimized loss function and set hyperparameter are used to train the network;Finally, the trained model is used to realize the rapid phase prediction of new metasurface structure.The application effectively solves the problems of phase jump processing difficulty in high degree of freedom metasurface design, and the lack of local and global feature collaborative modeling, realizes the rapid, high-precision and physically reasonable prediction of reflection phase in wide frequency band.

The present disclosure relates to the technical field of computer, and proposes a low-cost redundancy deployment method and scheduling device supporting availability guarantee, wherein a multi-dimensional fault feature model is constructed by collecting node and link failure probability, and a differentiated strategy set containing hot backup and cold backup is established; a recovery delay weighted expectation model is constructed based on backup priority and link state, and system-level recovery failure probability is quantified; a mixed integer linear programming model is used to globally optimize under the constraints of recovery delay threshold and availability probability, with the goal of minimizing the sum of static activation cost and dynamic recovery load, and generate a ternary mapping result; finally, the ternary mapping result is converted into node affinity, resource quota and routing rule of cloud-native cluster by a controller and automatically executed. The present disclosure realizes accurate balance between resource cost and availability guarantee, effectively avoids link failure risk, and realizes business-level recovery time guarantee.

The application relates to an intelligent electronic anti-shake method based on motion semantic recognition and adaptive multi-order filtering and an intelligent door lock, which comprises the following steps: S1, outputting a motion semantic label of a current frame according to the characteristic of angular velocity data; S2, inputting the angular velocity data into a fast response filter, a stable trend filter and a super stable filter in parallel, calculating corresponding fusion weights, and obtaining a smooth attitude quaternion of the current frame through ternary spherical linear interpolation; S3, constructing a space-time feature model, extracting node features by using an attention network of the space-time feature model, predicting a future attitude trajectory, performing prospective compensation correction, and obtaining a final compensation attitude quaternion; and S4, constructing a unified inverse mapping function, performing single interpolation resampling on an original image according to the unified inverse mapping function, and generating a stable output image frame.

The application belongs to the technical field of multimedia, and particularly relates to a multimedia content pushing optimization method and system based on artificial intelligence, which comprises the following steps: determining user characteristics based on user historical behavior information; comprising: obtaining historical browsing records, playing history, search history, collection records, like history and comment history of the user to combine as user historical behavior information; extracting content browsing preferences of the user from the user historical behavior information. The application determines the user characteristics by constructing a mapping from the user historical behavior information to the key behavior node and screening the preliminary label structure based on a preset evaluation score, converts the user behavior into a hierarchical clear and quantifiable feature model, overcomes the problem of insufficient recognition accuracy caused by the dependence of the prior art on simple label aggregation, and deeply mines the potential interest of the user to lay a foundation for subsequent accurate pushing.

The application relates to the technical field of transformers, and provides a transformer fault data sample generation method, which comprises the following steps: obtaining fault data of a transformer, sequentially collecting feature values from a value set of each feature of the fault data, obtaining temporary synthetic data according to the feature values, inputting the temporary synthetic data into a feature model to obtain predicted fault data output by the feature model, wherein the feature model is obtained by training a value set of other features of the fault data based on one feature of the fault data as a label, and new fault data samples are generated according to the predicted fault data. The transformer fault data sample generation method and device provided by the embodiment of the application learn the relationship between fault data features by establishing a model, and generate new fault data samples according to the learned feature model, thereby expanding the range of fault data.

The application discloses a gate real-time opening and closing strategy control system based on deep learning, comprising the following modules: a data acquisition module for generating a real-time water regime sequence; a feature modeling module for inputting the real-time water regime sequence into an improved Perceiver IO model to obtain a disturbance evolution feature vector sequence and an explanation field result; a trend risk calculation module for obtaining a control trend value and an oscillation risk value; an action generation module for generating a preliminary gate opening and closing action; a manifold analysis module for performing manifold curvature singularity dispersion analysis to obtain a water regime fluctuation intensity; an action adjustment module for adjusting the preliminary gate opening and closing action; and a control execution module for generating a gate opening and closing control instruction to perform opening and closing control on a gate actuator. The application realizes water regime disturbance modeling and gate intelligent control by combining the improved Perceiver IO model and the manifold curvature singularity dispersion analysis.

The present application belongs to the cross technical field of computer vision and intelligent transportation, and provides a video target dynamic identification and tracking method, aiming at solving the problem that the tracking precision, robustness and efficiency are difficult to be considered in traditional video target tracking due to target appearance change, scene interference and motion uncertainty, the method initializes the target in the video sequence and constructs a multi-granularity feature model fusing global and local features; for each subsequent frame, the global scene semantic information containing lane line topology structure is synchronously acquired, the constraint motion path hypothesis of the target is generated based on the path reasoning model through the traffic rules, and the candidate search area is demarcated accordingly; the candidate target is detected in the search area, the matching degree is calculated by using the multi-granularity feature model to determine the target position and update the state; the multi-granularity feature model is adaptively updated when the matching confidence is high, and the precision and robustness of long-time and stable tracking of the target in a dynamic complex scene are improved.