265 results about "Multi-label classification" patented technology

The invention discloses an intelligent monitoring management method and system based on archive digitization, and relates to the technical field of data management, and the method comprises the steps: collecting and preprocessing multi-source archive data, employing a multi-mode BERT model to carry out the feature fusion of different data sources, and generating a unified semantic representation; semantic labeling is performed on archive data through a multi-label classification model, a semantic graph of archive content is constructed by using a graph database, an association relationship between archives is represented, a semantic index tree is constructed based on the semantic graph, and rapid positioning and calling of the archive content are optimized; and recording the change of each file version, positioning the change position based on a semantic index tree, identifying the semantic change of the file through a semantic difference comparison algorithm, recording hash, carrying out granularity division on the file content through the semantic boundary of each level of node in the index tree, and generating a user access strategy. According to the invention, dynamic perception and risk early warning of user behaviors are realized, and the intellectualization and safety of the archive management system are effectively improved.

The invention discloses an incomplete multi-view multi-label data classification method based on semantic enhancement and pseudo-label uncertainty perception, and the method comprises the steps: employing a dual-channel feature extraction and decoupling module to obtain the shared semantic representation and specific representation of each view in each sample for a constructed incomplete multi-view multi-label data classification network model; performing cross-view fusion on the shared semantic characterization and the specific characterization, obtaining a unified shared characterization and a unified specific characterization corresponding to each sample, performing feature fusion, obtaining a fusion characterization of each sample, inputting the fusion characterization of the sample output by the dual-channel feature extraction and decoupling module into a classifier for multi-label prediction, and performing multi-label prediction on the fusion characterization of the sample. Therefore, a multi-label classification prediction result is obtained, and model training is carried out based on a total contrast learning loss function and a joint supervision classification loss function. According to the method, the classification performance and the model robustness on incomplete multi-view multi-label data are remarkably improved through training learning under the guidance of semantic enhancement and uncertainty.

The invention relates to the technical field of artificial intelligence auxiliary medical treatment, and discloses a child tooth trauma intelligent grading and dynamic processing decision auxiliary system. A trauma identification module; an injury condition grading module; a processing suggestion generation module; a prognosis risk dynamic evaluation module; and a man-machine interaction module. The method comprises the following steps: acquiring an image containing child tooth trauma information and clinical data; utilizing a multi-label classification model to identify a trauma type; calculating a comprehensive injury condition score through a hierarchical multi-modal scoring network and determining an injury condition grade; calling a knowledge graph and a rule engine to generate and dynamically correct a processing suggestion; and continuously updating the prognosis risk assessment result based on the time sequence review data. The invention aims to solve the problems of strong evaluation subjectivity and complex decision in the diagnosis and treatment of the tooth trauma of children, provides standardized and intelligent decision assistance for clinic, and realizes closed-loop prognosis management.

The invention relates to a semi-supervised semantic segmentation method for multi-scale patch classification for sea target identification, and belongs to the technical field of computer vision. Inputting the marked data and the unmarked data into a teacher-student framework to carry out weak disturbance and strong disturbance, and then carrying out prediction; the teacher MPMC module performs multi-scale pooling operation on the extracted features and connects the features after multi-scale pooling; classifying the featured receptive field patches through a convolutional network and a linear layer, and calculating an adaptive weight according to a classification result; the self-adaptive weight sum is used for adjusting the teacher model and the student model; calculating the supervision loss between the teacher model prediction result and the student model prediction result, and calculating the multi-label classification supervision loss of the MPMC to the marked data; updating parameters of the student model through an optimization algorithm; and the student segmentation network predicts the input image to obtain a final prediction result. According to the method, the data annotation requirement is reduced, the segmentation precision is improved, and the robustness and generalization ability of the model are enhanced.

The invention discloses a processing method and system for rejecting and hanging work order data, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining the rejecting and hanging work order data and a real label corresponding to the rejecting and hanging work order data; inputting the rejected work order data into a trained basic text model to obtain a reasoning result, the reasoning result comprising a prediction label and a confidence coefficient; screening out the rejected work order data of which the confidence coefficient is smaller than a preset value and the predicted tag is inconsistent with the real tag from the reasoning result as low-quality data; determining a quality problem type of the low-quality data based on a quality problem determination rule; performing iterative optimization on the basic text model by adopting a corresponding optimization strategy based on the quality problem type to obtain a multi-label classification model; and inputting the to-be-improved work order rejecting and hanging data into the multi-label classification model to obtain an optimal reasoning result, thereby facilitating solving the problem that the reasoning result of the work order rejecting and hanging data cannot be accurately obtained in the prior art.

The invention discloses an OCR (Optical Character Recognition) optimization method based on multi-label classification and active learning, and relates to the technical field of image recognizing.The method comprises the steps that an original image-text image is obtained and preprocessed, problem type classification is performed on problems existing in the preprocessed image based on a multi-label classification model, corresponding image problem types are obtained, and meanwhile, priorities are generated; performing targeted enhancement on the image according to the priority of the image problem classification result, repairing various problems existing in the image, and identifying the repaired image to obtain a structured character identification result list; and verifying an identification result, screening out difficult sample data, entering an active learning link according to the number of the sample data, and optimizing a multi-label classification model at the same time, thereby realizing dynamic cooperation of image quality improvement and OCR performance optimization. The problem that the accuracy of OCR recognition is reduced due to the multi-source quality degradation problem of the image in the prior art is effectively solved.

The invention relates to the technical field of artificial intelligence, and provides a protein function prediction method and device based on multi-modal protein data, and the method comprises the steps: obtaining protein multi-source data, carrying out the feature extraction of a protein sequence in the protein multi-source data, and obtaining a protein sequence feature; constructing a heterogeneous graph based on the protein multi-source data; performing feature coding on the heterogeneous graph by adopting a graph attention mechanism to obtain protein graph features; performing multi-modal fusion on the protein sequence features and the protein map features by adopting a gating fusion mechanism to obtain fusion features; and performing multi-label classification prediction based on the fusion features to obtain a protein function annotation result. The accuracy and robustness of protein function prediction can be improved, and the problems that in the prior art, multi-source protein data cannot be effectively integrated, and the method is sensitive to data noise are solved.

The invention relates to the technical field of wind turbine generator fault early warning, and discloses a wind turbine generator fault early warning method and system based on multi-modal data fusion, and the method comprises the steps: collecting the data of a multi-modal sensor, and carrying out the time-space alignment preprocessing; multi-modal features are extracted through variational mode decomposition, STL decomposition and other methods, and cross-modal fusion is achieved through dimension adaptive projection and a multi-head attention mechanism; calculating a dynamic weight based on three factors of data quality, fault type correlation and information gain, and carrying out weighted fusion; constructing a dynamic unit topological graph, and capturing cross-unit association features by using a space-time diagram convolutional network; long-time early warning with confidence is realized through double-branch gating fusion in combination with a Bayesian neural network; a multi-label classification identification multi-fault mode is adopted, and an operation and maintenance decision is optimized through an adaptive large neighborhood search algorithm. According to the method, the long early warning window of the offshore wind turbine generator can be realized, and uncertainty quantification and intelligent operation and maintenance decision support are provided.

The invention discloses a multi-label classification method and system for an order sending scene, and the method comprises the steps: carrying out the feature extraction of an input work order text, and generating a global text semantic representation; based on a pre-constructed semantic embedding vector of each preset service label, performing label knowledge enhancement on the global text semantic representation to realize alignment of the text representation and the label semantics in a semantic space; based on the aligned text semantic representation, calculating an original confidence coefficient corresponding to each preset service tag through a main classifier; for each preset service label and the work order text, dynamically generating a classification threshold value corresponding to the service label; and judging the original confidence degree based on the classification threshold, and outputting a multi-label classification result corresponding to the work order text to realize a differentiated order dispatching decision. Therefore, by means of the personalized classification threshold value generated dynamically, the problems of misdispatch and missed dispatch caused by the fixed threshold value in the dispatch scene are effectively solved, and the accuracy of multi-label classification is remarkably improved.

The invention discloses a multi-layer circuit board quality inspection method and system based on machine learning, and the method comprises the steps: carrying out the time-space registration of collected multi-source data through an adaptive weighted fusion algorithm, and generating a multi-mode quality inspection data set containing a line topological structure and material characteristics; outputting a fused circuit board defect sensitive feature vector set by using a pre-trained nested attention deep learning model based on the multi-modal quality inspection data set; inputting the defect sensitive feature vector set into a twin network architecture, positioning a potential defect area through a dynamic anchor frame generation mechanism, carrying out multi-label classification on defect types in combination with a Bayesian probability model, and synchronously introducing a defect severity evaluation module to quantify the influence degree of defects on circuit performance, and outputting a detection result containing the defect position type and severity. According to the embodiment of the invention, the collaborative judgment of the type, position and severity of the defect can be realized, and the detection precision and generalization capability of the defect of the multilayer circuit board are improved.

Methods and systems for classification include performing multi-label classification on an input using a trained model to generate classification outputs corresponding to respective labels. The classification outputs are fused to generate a joint opinion. It is determined that the input is out of distribution as compared to a training dataset of the trained model based on a joint belief of the joint opinion. An action is performed responsive to the determination that the input is out of distribution.

The invention discloses a composite material reflectivity spectral information classification method and system based on a PCA-SVM algorithm. The method comprises the following steps: collecting reflectivity spectral data of a composite material sample; preprocessing data to eliminate measurement deviation and unify numerical scale; carrying out dimensionality reduction on the preprocessed high-dimensional spectral data through principal component analysis, and extracting feature components retaining main variance information; based on dimension reduction features, a support vector machine is adopted to construct a multi-label classification model according to a'one-to-other 'strategy; predicting the test sample, and generating a multi-label classification result through probability output and threshold processing; and analyzing the classification performance by using the multi-label evaluation index. The data processing module of the system executes preprocessing, dimension reduction, modeling, prediction and evaluation operations. The method is suitable for lossless identification of multi-component composite samples, both interpretability and identification precision are considered, the performance bottleneck of a traditional method under the conditions of feature overlapping, insufficient samples and the like is effectively overcome, and efficient and accurate classification of the composite materials is achieved.

The invention relates to the technical field of video quality diagnosis, in particular to a video quality diagnosis method and system based on large model extensible classification, and the method comprises the steps: carrying out the sliding window sampling of an input video stream, recognizing and positioning an abnormal frame, and obtaining a plurality of image frames; respectively extracting visual features corresponding to the image frames and adding time codes and label codes; fusing time sequence information of historical frames through time coding and label coding to generate global time sequence characteristics; based on predefined abnormal category label text features, a label semantic vector set is generated and spliced, a query vector is generated through a full connection layer, an attention mechanism is utilized to match video features and label features, and a multi-label classification result is output; dynamically triggering an optimization process according to the abnormal confidence coefficient, and generating new tag text features through Few-shot sample input and low-rank adaptation fine tuning; and adding the new label text features into the label semantic vector set. The accuracy of a classification result with a small sample size is improved.

A system and method for dynamically optimizing large language model (LLM) inference by selectively deactivating layers based on query complexity. A multi-label classifier is trained on diverse user queries and their optimal layer configurations. During inference, the classifier analyzes incoming queries to predict which LLM layers can be safely deactivated without compromising output quality. The system processes user queries through the LLM with the predicted layer configuration, reducing computational resources while maintaining accuracy. A database stores historical queries, layer configurations, and performance metrics for continuous system improvement.

The invention relates to the technical field of computer vision, in particular to a tongue picture multi-label multi-task automatic classification method and system.The method comprises the steps that tongue picture images are obtained to construct an original data set, and the tongue picture images in the original data set are labeled; performing preprocessing based on the obtained original data set, including introducing an illumination compensation factor to perform adaptive brightness normalization on the tongue picture image; the pre-processed tongue picture image is used as input to construct a classification model, multi-level feature extraction is carried out, multi-task feature decoupling is carried out by using the extracted multi-level features, multi-label classification and loss function optimization are carried out based on a feature decoupling result, and the trained classification model is optimized by using a Lion optimizer. According to the method, feature maps of different levels are fused in a cross-stage manner, multi-scale capture of tiny lesions in the tongue picture image is realized, and the detection capability of the model on features such as tongue color, tongue coating color, tongue shape and tongue coating quality is remarkably improved.

The invention discloses a pedestrian attribute identification method based on occlusion recovery and visual angle fusion. A used model comprises a backbone network, a self-adaptive shielding module, a visual angle perception feature fusion module, a missing region feature reconstruction module and a multi-label classifier. Extracting initial pedestrian features from the pedestrian image through a backbone network; the self-adaptive shielding module extracts pure pedestrian features by using the initial pedestrian features and the random mask; a visual angle perception feature fusion module extracts visual angle common semantic features and enhances the pure pedestrian features to obtain the pure pedestrian features with visual angle common semantic enhancement; a missing region feature reconstruction module performs refined feature reconstruction on the missing region to obtain missing region reconstruction features; and after the pure pedestrian features, the pure pedestrian features with the view angle common semantic enhancement and the trend region reconstruction features are fused, pedestrian attributes are identified through a multi-label classifier. According to the method, the recognition precision and robustness of the model in complex scenes such as shielding and multi-view angles are improved.

A method for identification of land improvements in a given parcel of land that includes generating a first training data set by clipping a large image of a parcel of land into individual parcel images each including images of improvements, and each improvement being labeled with an improvement type, providing the individual parcel images to a first classification model, training the first classification model based on the individual parcel images to identify unlabeled improvements in a parcel image and to obtain a multi-label classifier, generating a second training data set of images, and training a second, semantic segmentation model based on the second training data set.

The invention relates to a BERT and multi-module fusion-based track circuit fault label classification method and device and a medium, and the method comprises the steps: obtaining a to-be-classified track circuit fault text, and carrying out the cleaning and standardization processing of the to-be-classified track circuit fault text; performing word segmentation and vectorization processing on the track circuit fault text to generate a lexical symbol sequence conforming to BERT encoder input specifications; extracting a context semantic vector representation of the input lexical symbol sequence through a BERT encoder; performing feature extraction through a TextCNN module and a self-attention module which are parallel to obtain local features and weighted global features, and splicing the local features and the weighted global features to obtain a comprehensive feature vector; and synchronously generating a multi-label classification result of a fault phenomenon, a fault reason and a solution measure corresponding to the current fault text through a multi-label classification head based on the comprehensive feature vector. Compared with the prior art, the method has the advantages of high discrimination, high accuracy, high practicability and the like.

The invention discloses a multi-label interference identification method and device based on dual-domain asymmetric feature fusion, and the method comprises the steps: enabling a received interference signal to generate two types of time-frequency images through short-time Fourier transform and continuous wavelet transform, converting the two types of time-frequency images into RGB images, and inputting the RGB images into a dual-branch AsymResNet18FPN network to extract multi-scale features; cross-domain feature adaptive fusion is realized through channel splicing and attention weight generated by MLP; and finally, outputting an interference type combination by the multi-label classifier. According to the method, the complementarity of double-domain features is fully utilized, the direction perception and multi-scale modeling capability is enhanced, the problem of feature submerging under the low interference-to-signal ratio is effectively relieved, synchronous recognition of multiple composite interferences is supported, and the recognition precision and robustness in a complex electromagnetic environment are remarkably improved.

The invention discloses a multi-mode resume layout self-adaptive analysis method and system based on a large model, and particularly relates to the technical field of self-adaptive analys.The multi-mode resume layout self-adaptive analysis method comprises the steps that firstly, resume files of multiple formats uploaded by a user are converted into standard image formats in a unified mode, and edge density calculation and language recognition are conducted on images; and judging whether to perform image enhancement processing or not according to the calculated image enhancement coefficient. Then, text information is extracted through the OCR technology, multi-modal representation is constructed through fusion of a visual feature extraction model and a large language model, and paragraph positions are corrected in combination with a diffusion layout model; the system further adopts a multi-label classification model to identify a functional region, constructs a similar sample set by matching historical resumes, calculates the weight of each region based on HR attention and content deviation, performs sorting and length optimization on resume contents, and finally generates a personalized resume structure which is clear in structure and prominent in expression. The system is suitable for resume analysis and content recommendation in a complex layout and multi-language environment.

The invention discloses an automatic cue word optimization method, system and equipment based on large language model strategy priori, and belongs to the technical field of artificial intelligence and natural language processing. The method comprises the following steps: constructing a double-model collaborative architecture which takes a large language model as a strategy prior network and takes a target small model as an evaluation environment; analyzing error cases of the target model on the verification set by using a large language model, and generating candidate rule modification actions and confidence thereof; taking the confidence coefficient as a prior probability, fusing the confidence coefficient into a Monte Carlo tree search process based on a PUCT formula, and dynamically balancing semantic intuition and actual measurement feedback; adopting a dynamic funnel evaluation and safe rollback mechanism to implement zero-tolerance pruning on modification causing performance degradation; and in combination with incremental optimization and a historical freezing strategy, serially optimizing cue words in a multi-label classification scene. According to the method, the high-quality and conflict-free prompt word rule set can be efficiently and automatically generated, and the performance of a small model on a specific task is remarkably improved.

The invention discloses a drainage pipe defect identification and positioning method and system based on a class activation diagram, and belongs to the field of drainage pipe network defect detection. The method comprises the following steps: processing data set label description and learnable category prompt through a text encoder, generating label text features and category prompt features, and extracting local features and global features of a drainage pipe image by using a visual encoder; fusing the category prompt features and the image local features, and combining the image global features and the label text features to obtain global fusion category prompt features; calculating the similarity between the global fusion category prompt feature and the converted image local feature, and generating a multi-label classification result; model parameters (frozen text encoder) are optimized by prompting enhanced training. And processing a classification result based on a Grad-CAM technology to generate a class activation diagram, and outputting a visual heat map through interpolation and mapping to position a defect position. Accurate identification and visual positioning of a defect area are realized by fusing a class activation graph technology.

The invention discloses a multi-label field adaptive method based on prompt driving, which belongs to the technical field of computer vision and comprises the following steps: generating multi-dimensional semantic text description containing visual attributes, semantic levels and category names of common scenes for each type; embedding the semantic text description into an optimizable category vector, and embedding the optimized category vector into a CLIP text prompt; extracting and projecting multi-layer style statistical characteristics of the image, and synchronously realizing image-text cross-modal alignment and source-target domain distribution alignment in a frame through a cross-domain style mapping network and various alignment losses; semantic priori knowledge of CLIP is combined with a label co-occurrence mode of a source domain, so that the model can sense and adapt to a possibly changed label dependency relationship in a target domain; performing semantic propagation on label embedding by using a graph convolutional network; the whole system is jointly optimized through a multi-task loss function, and end-to-end cross-domain multi-label classification is achieved.

The invention discloses a traditional Chinese medicine tongue picture intelligent analysis method based on YOLO target detection and multi-label classification, and relates to the technical field of computer vision. The method comprises the following steps of: firstly, importing an original tongue picture color image into a tongue surface area and sublingual area detection model based on a YOLOv11 target detection network to obtain a tongue surface area and sublingual area detection result; the method comprises the following steps: cutting an original image to obtain a regional image of a target object, carrying out zooming processing to obtain a new regional image with a standard size, importing the new image into a tongue picture multi-label classification model which is trained for the target object and is based on a Resnet50 network, and obtaining a tongue picture multi-label classification result of the target object. And finally, summarizing to obtain a traditional Chinese medicine tongue picture intelligent analysis result corresponding to the original image, thereby effectively improving the analysis result precision through lingual surface / sublingual region separation analysis and cutting operation, and reducing the model coupling degree to a certain extent.

The invention discloses a sliding hybrid model construction method, and particularly relates to the technical field of natural language processing and artificial intelligence, and the method specifically comprises the following steps: S1, voice-to-text and small model prediction; s2, performing Flag preliminary judgment and output; s3, threshold table judgment and model selection; and S4, building and predicting a large model prompt. The invention relates to a sliding hybrid model construction method, aims to solve the problems of low efficiency, poor accuracy, large resource consumption and the like in related business applications, and provides a sliding hybrid model construction method through construction of a long-tail intention threshold table, information extraction and classification based on prompt, 'prior information + PR curve 'threshold analysis and the like. The long-tail intention is quickly processed, the multi-label classification accuracy is improved, and the model performance is stabilized. The application effect is good in scenes such as automobile sales and after-sales, an efficient and intelligent solution is provided for multi-label classification and related services, and user experience and enterprise benefits are effectively improved.

The invention belongs to the technical field of large model application, and relates to an enterprise technology demand prediction method and system based on large model driving, and the method comprises four parts: enterprise problem deep analysis, problem technology essence identification, technology knowledge base intelligent retrieval and multi-granularity technology demand prediction. In the deep analysis of enterprise problems, the problem description is subjected to structured processing, an entity relationship graph is constructed, and domain knowledge is aligned. The problem technology essence identification utilizes large model semantic understanding to analyze core challenges, and classifies technical problems through a multi-label classification model. And the technical knowledge base intelligently retrieves and calculates the similarity between the problem semantic vector and the knowledge graph, fuses multi-source data expansion, and screens candidate technologies according to the technology maturity. And performing multi-granularity technology demand prediction analysis on the core technology field and the subdivision direction to form a multi-level demand result. According to the invention, through systematic analysis and intelligent prediction, enterprise technology demands are accurately identified, and the method is suitable for scenes such as industry-university-research cooperation, technology transfer and innovative resource configuration.

The invention discloses a bridge concrete apparent damage identification method and system based on texture analysis, and belongs to the crossing field of bridge engineering detection and computer vision technology, and the method comprises the steps: employing an unmanned plane to collect a bridge member surface image, and generating an orthoimage; extracting contrast, correlation and entropy features based on the gray-level co-occurrence matrix; extracting histogram features based on a local binary pattern; detecting a corrosion dialysis area based on an HSV color space; fusing the multi-dimensional features, inputting the fused multi-dimensional features into a multi-label classifier, and outputting pixel-level segmentation masks of various damages such as cracks, spalling, exposed reinforcement corrosion and dialysis at the same time; according to the method, the complementary texture features are fused to realize simultaneous identification of multiple types of damages, a pixel-level segmentation result is output, and an evaluation report conforming to engineering specifications is automatically generated.

The invention belongs to the field of multimedia information intelligent processing and classification application, and relates to a multimedia field label information labeling method based on TSK rule migration. The method comprises three parts of label set division, rule migration modeling and multi-label classification. The method comprises the following steps: firstly, constructing a label separation mechanism according to label distribution characteristics of multimedia multi-label data, and dividing a label set into conventional labels and scarce labels; secondly, constructing a rule migration TSK fuzzy system; the system comprises a combined If-part and a migration-driven ten-part, not only can realize feature-label reasoning modeling, but also can establish relevance between conventional labels and scarce labels according to label distribution differences. And finally, on the basis of RT-TSK-FS, a multi-label classification method based on rule migration is provided, and sharing and supplement of knowledge among different labels are realized through a rule migration mechanism, so that the prediction performance of multimedia label information is improved.

The invention discloses a mobile application privacy policy compliance detection method based on a pre-training model, and aims to realize efficient and intelligent privacy policy compliance automatic detection. The method comprises the following steps: firstly, constructing a hierarchical privacy policy compliance detection index system according to domestic related laws and regulations and standards; secondly, collecting an original text of the privacy policy through a web crawler technology, and constructing an unlabeled corpus after cleaning and structured processing; thirdly, constructing a multi-label classification data set based on a mode of combining a large language model and manual review; mapping the text and the label to a unified semantic vector space by adopting a text-label joint embedding strategy, and inputting the text and the label into a multi-granularity classification model; according to the model, on the basis of an ERNIE pre-training model, context feature enhancement and deep semantic interaction are realized through a bidirectional long-short-term memory network, a self-attention mechanism and a text-label cross attention mechanism, so that the multi-label classification performance is remarkably improved; finally, according to a label prediction result output by the model and a preset index system, compliance judgment is automatically completed, and a structured detection report is generated. According to the invention, the automation degree and efficiency of privacy policy compliance detection are effectively improved.

A logfile recommender service incorporates trained machine learning models to predict recommended logfile types for support tickets. The machine learning models are trained using data from past support tickets such as titles, problem descriptions, and communication records. After receiving a request to generate a logfile recommendation for a given support ticket, the logfile recommender service first calls an attachment prediction machine learning model (e.g., a binary classification model) to predict whether logfiles are relevant to the support ticket. If so, the logfile recommender service calls an attachment recommendation machine learning model (e.g., a multi-label classification model) to generate a list of recommended logfile types for the support ticket, which is output by the logfile recommender service as the logfile recommendation. Logfiles of the recommended types can then be attached to the support ticket to facilitate and expedite resolution of the support ticket.