36 results about "Final Labeling" patented technology

The invention discloses an automatic label labeling and classifying method and system oriented to unstructured system documents, and relates to the technical field of artificial intelligence. The method comprises the steps that semantic structure pre-analysis is conducted on an original system text, and a system semantic structure tree is constructed; establishing a system semantic enhancement vector space based on the semantic units and the logic relationship thereof; performing semantic deconstruction on the preset tag and extracting a feature vector; realizing cross-space semantic matching of the document and the tag through a system semantic attention mechanism; a confidence evaluation module is introduced to screen high-confidence labels from the three dimensions of structural integrity, coverage and logic consistency; and outputting a final label and a score through semantic conflict detection and resolution. According to the method, the problems that in the prior art, unstructured system text labeling accuracy is low and large-scale labeling samples are dependent on polysemy ambiguity, high context dependency, complex semantic structure and the like are solved, and labeling accuracy and robustness are remarkably improved.

The invention provides a public opinion video tag aggregation method and system based on artificial intelligence, and relates to the technical field of artificial intelligence. Comprising the following steps: acquiring pictures and text information in a short video, and performing semantic alignment; different large language models are adopted to generate preliminary labels for the pictures and the text information after semantic alignment; clustering the pictures and the texts after semantic alignment to obtain clusters; calculating the labeling probability of each primary label type in the current cluster by each large language model, and selecting the primary label with the highest probability sum as a clustering label of the current cluster; calculating the reliability weight of each large language model in the current cluster based on the clustering label of the current cluster; and based on the reliability weight, calculating the weighted support degree of all the large language models to different preliminary label types of each piece of data in the current cluster, calculating the weighted label of the current data, and further determining a final label. According to the method, the condition of few labels or no labels can be effectively processed, and the manual workload is greatly reduced.

The invention relates to the field of electronic design automation, provides a method for semi-automatically processing an EDA (Electronic Design Automation) project file and generating a YOLO format annotation data set, and aims to solve the technical problems of low efficiency and poor precision of the traditional pure manual data annotation data set. According to the main scheme, the method comprises the steps of reading an engineering file of a jia-creative EDA project, analyzing and extracting element information, generating a standardized CSV data format, converting the standardized CSV data format into a label in a YOLO format, calculating normalized bounding box coordinates of elements, and recognizing special original elements and supplementing labeling information in combination with a template matching technology; drawing a bounding box on the schematic diagram image according to a YOLO label, highlighting different element types by using different colors, and generating a visual annotation image; revising and combining the newly generated YOLO label and a preliminary YOLO label generated by the AI model to obtain a label set; then importing into Labelimg for manual correction, and generating a final label set; and training a target detection model by using the final label set, and improving the pre-labeling precision through data set expansion iteration.

The invention discloses a quality inspection and labeling system for bid invitation file generation based on deep learning, and the system comprises a data standardization module which is used for collecting items and historical texts and executing data standardization, and generating a standardized text sequence; the semantic coding and candidate generation module is used for tensorizing the standardized text sequence and generating a candidate text of the bid invitation file; the text decoding module is used for executing limited CRF decoding on the candidate text and outputting an aligned labeling result file; the hypergraph labeling module is used for constructing a hypergraph and globally decoding the hypergraph to generate a multi-level labeling result; the selective re-checking module is used for calculating uncertainty, triggering manual re-checking and outputting a confirmed labeling result and an uncertainty record; the consistency verification module is used for verifying hierarchical consistency to form a final labeling result; and the structured export module is used for exporting the bid invitation file finished product text and the structured annotation data. According to the invention, quality inspection and labeling of bid invitation file generation are realized.

The invention provides a multi-label path reasoning method and system, a medium and a terminal. The method comprises the following steps: constructing a label system; dynamically updating a keyword set in the tag system; and based on the tag system and the keyword of the tag node, allocating a final tag to the to-be-classified text. According to the multi-label path reasoning method and system, the medium and the terminal, multi-label attribution can be efficiently and accurately completed under a complex label system through label system structured grouping, upper and lower layer path collaborative judgment and an embedded semantic fusion and path scoring mechanism.

The invention relates to an e-commerce scene-oriented commodity label generation method and device, and a storage medium, and relates to the technical field of artificial intelligence. The method comprises the steps of obtaining native text information and an image set of a target commodity; performing information identification and extraction on the image set, converting related visual data into structured text description data, and fusing the structured text description data into a to-be-processed text corpus with multi-dimensional information; dynamically generating a theme label structure for the target commodity; performing text segmentation on the to-be-processed text corpus according to the topic tag structure to obtain grouped texts corresponding to different topic tags, and feeding back each grouped text to a corresponding topic expert sub-model to perform candidate tag extraction; and semantic fusion is carried out, and a fused label set is used as a final label of the target commodity. According to the method, the integrity, accuracy and robustness of commodity label generation in an e-commerce scene are remarkably improved, and the limitation of dependence on manpower or a fixed label library is reduced.

The invention discloses a visual segmentation large model data automatic annotation method combined with image prompt, which comprises the following steps: extracting deep features of a prompt image and a to-be-annotated image by adopting a shared visual coding technology of deep learning, and carrying out cross-image feature matching and similarity calculation. And positioning a high-response area which is consistent with a prompt target in semantics in the to-be-labeled image, and automatically screening out an accurate prompt point set from the high-response area. Furthermore, based on the point set, the visual segmentation large model is guided to complete accurate segmentation and mask generation of the target, so that full-process automation from image prompting to final labeling is realized, and labeling efficiency and consistency are remarkably improved. In this way, the bottleneck problems that in the prior art, an automatic labeling method is insufficient in universality and the application of a segmentation large model must depend on human interaction are solved, and an effective technical approach is provided for large-scale, low-cost and high-quality visual segmentation data production.

The application discloses a text data intelligent labeling system and method based on a large language model, and belongs to the technical field of natural language processing. The system and method are characterized in that a complete modification operation sequence of a large language model pre-labeling result of an artificial labeler is collected to construct an interactive log containing an original text, a pre-labeling result, a modification action and a final labeling result; the distribution law of the modification operation in the log is analyzed, a multi-level defect classification system is constructed in combination with the association between a text semantic feature and a modification type, and the influence weight of each defect is quantified through the construction of a defect association network atlas; feature samples are extracted for various defects, a defect training set is constructed, and a semantic enhancement instruction set is generated; meanwhile, resources are distributed according to the defect influence weight, the large language model is subjected to staged parameter calibration, and the corresponding calibrated model parameters are dynamically called for pre-labeling in subsequent labeling.

Labelling in machine learning is disclosed. Assistant models are trained to generate the labelling behavior of corresponding labelers. When an observation is received, the label generated by the assistant model is compared to the user-generated label. If the labels are different, an ensemble operation is performed wherein multiple assistant models, each configured to mimic the labelling behavior of a different user, are used to resolve the conflict and determine a final label for the observation. Labels generated by the ensemble operation may be incorporated into retraining operations performed to retrain the assistant models.

The invention belongs to the technical field of astronomical image data processing, particularly relates to an astronomical image data set labeling method, and aims to solve the problem that existing single threshold segmentation-based labeling and star map projection-based labeling are difficult to meet the requirement of a depth model for a high-quality data set. According to the method, on the basis of threshold segmentation labeling and star map projection labeling, a first labeling result and a difference set are adopted as a labeling result A for an observation image large-probability detectable target, a second labeling result is adopted as a labeling result B for an observation image small-probability detectable target, and the labeling result A and the labeling result B are integrated to obtain a labeling result A and a labeling result B; and obtaining a final labeling result. According to the method, the integrity of target labeling can be improved, the probability of missing labeling and wrong labeling can be reduced, the actual fitting degree of the labeling result and the image can be met, unification of integrity and accuracy is achieved, and the depth model performance can be improved through the manufactured data set.

The application discloses a 3D target labeling method and system fusing a visual large model and point cloud projection, relates to the technical field of 3D target automatic labeling, and comprises the following steps: performing pixel-level semantic segmentation on a group of collected images and point clouds through a visual large model to obtain a segmentation graph of a target pixel region and a semantic label of each pixel; generating point clouds with semantic labels by using projection of the point clouds to an image plane and semantic mapping; clustering and processing foreground point clouds belonging to the same semantic label to obtain point cloud clusters belonging to the same semantic label; traversing the point cloud clusters, fitting a 3D bounding box, combining parameters of the fitted 3D target box with the same semantic label, generating 3D target labeling post-processing of the same semantic label, and obtaining a final labeling result file; and the 3D target labeling method realizes efficient and accurate 3D target automatic labeling, significantly reduces labeling cost, and improves the quality and consistency of labeled data.

In some aspects, a computing device can receive, at a data processing system, a set of utterances for training a named entity recognizer or for inference with a named entity recognizer to assign a label to each token segment in the set of utterances. The computing device can determine a length of each utterance in the set and, when the length of the utterance exceeds a predetermined threshold of token segments: divide the utterance into a plurality of overlapping token segment chunks; assign a label and a confidence score to each token segment in the token chunk; determine a final label and an associated confidence score for each token segment chunk by merging two confidence scores; determine a final annotated label for the utterance based at least on merging the two confidence scores; and store the final annotated label in a memory.

The invention discloses a data security labeling method and system based on privacy computing, and relates to the technical field of data processing, and the method comprises the steps: receiving a labeling task request, decomposing the labeling task request into a plurality of sub-tasks based on a task analysis model, and classifying and routing the sub-tasks to a federated learning space, a trusted execution environment (TEE) space and a secure multi-party computing (SMC) space; in the federated learning space, executing a first type of subtasks based on a teacher-student model architecture; and executing the second type of subtasks in the TEE space, executing the third type of subtasks in the SMC space, and aggregating subtask processing results to generate a final labeling result. The technical problem that in the prior art, a complex data labeling task cannot be executed safely and with high quality under the condition that multi-mechanism data does not go out of a domain is solved, and the purpose that multi-mechanism data labeling is completed through cooperation of federated learning, a trusted execution environment and safe multi-party calculation is achieved. The technical effect of safe and high-quality data annotation processing on the premise that multi-mechanism data is not out of a domain is achieved.

The application provides a multi-modal data labeling method and device and a storage medium, and relates to the field of automatic driving. The method comprises the following steps: synchronously collecting multi-modal data and obtaining environment parameters; based on the environment parameters, obtaining modal reliabilities, judging a modal with the highest modal reliability as a reliable modal, and judging other modals as unreliable modals; a pre-labeling result of the reliable modal is a final labeling result of the reliable modal; the pre-labeling result of the reliable modal is transformed to obtain a projection result of the pre-labeling result of the reliable modal on the unreliable modals; the projection result of the reliable modals and the pre-labeling result of the unreliable modals are input into a target association module to output an intersection-over-union ratio; and the pre-labeling result of the unreliable modals is corrected according to the intersection-over-union ratio to obtain a final labeling result of the unreliable modals. The method effectively solves the problems of poor environment adaptability and insufficient multi-modal data fusion precision in the prior art.

The invention relates to the technical field of semiconductor manufacturing quality control, in particular to a space-time coupling virtual measurement intelligent labeling method based on feature space geometry, which comprises the following steps: S1, acquiring multi-dimensional sensor time sequence data of semiconductor manufacturing equipment as a labeling sample set, and performing manifold value evaluation on the labeling sample set; s2, establishing an adaptive time weight model based on data complexity, establishing a dynamic space weight model based on local sparseness based on an adaptive kernel function, combining the adaptive time weight model, the dynamic space weight model and manifold value evaluation, and defining a time-space coupling value score; s3, taking the time-space coupling value score as an input, and establishing a three-layer optimization architecture to realize a globally optimal sample combination; s4, combining the samples selected by each cluster to form a final label set, performing quality evaluation on the final label set, and generating standardized label output according to a quality evaluation result; and an accurate value basis is provided for subsequent intelligent selection.

The application discloses a kind of road scene automatic labeling method, it is related to the field of automatic driving, comprising: obtaining multimodal sensor data, time synchronization camera image and laser radar point cloud;The multimodal sensor data is preprocessed;The camera image and laser radar point cloud after pre-processing are input into pre-labeling model, and pre-labeling result containing target category, three-dimensional space information and uncertainty score is generated;According to the uncertainty score, the pre-labeling result is sorted to obtain the final labeling result.High-efficiency multi-sensor fusion is realized, and more rich labeling of semantic and geometric information is generated;Through uncertainty quantification, the labeling efficiency is greatly improved, and the labeling quality is guaranteed.

The application discloses a kind of intelligent labeling method, equipment and medium of text data, belong to electric digital data processing technical field, comprising: obtaining the label ontology of source field and target field;Based on the concept granularity, semantic boundary and axiom form between the two, construct the semantic alignment graph of cross ontology;According to the formal semantic equivalence determination of the logic rule of the graph to source field, generate the translated logic rule suitable for target field;After multi-label prediction is carried out to target text, the consistency check is carried out to candidate label combination using translation rule, and the final labeling result is output.The application solves the logic rule invalidation problem caused by label system isomerism and semantic boundary dynamic change in cross-domain scene by constructing the migration framework based on formal ontology semantic alignment and hierarchical translation verification, realizes the semantic fidelity and structure consistent translation of logic rule between heterogeneous knowledge system, effectively avoids semantic drift and conflict misjudgment.

The embodiment of the invention provides a tag generation method and device, equipment, a storage medium and a product. The tag generation method comprises the following steps: performing multi-round iteration on a tag set of a full-quantity file set comprising a plurality of files; in each round of iteration, the fitness of the label set is calculated, the label set is adjusted according to the fitness, and a final label set is output until a preset iteration termination condition is met; the fitness is calculated according to evaluation comprehensive parameters, and the evaluation comprehensive parameters at least comprise one of semantic relevancy between each file and a label set, relevancy among all files under the same label and repetition among all files under the same label; file preference dynamic vectors driven by importance and / or user access behavior of each file can also be selectively incorporated. According to the embodiment of the invention, the tag set is optimized through multiple rounds of iteration, and the fitness is calculated in combination with the related core parameters between the files and the tags and the selectable user related factors, so that intelligent tag generation of massive files is realized, and the file management retrieval efficiency and the service competitiveness are effectively improved.

The application discloses a web-based continuous frame point cloud data labeling method and device, and the method comprises the following steps: obtaining a labeling request, responding to the labeling request, downloading continuous frame point cloud data to be labeled and corresponding 2D image data and storing the same in a local database; obtaining an initial labeling box of an object in a frame of point cloud data in the continuous frame point cloud data, automatically calculating a final labeling box based on the initial labeling box; and mapping the final labeling box to the 2D image data. The application can realize multi-frame big data loading and automatic convergence of the labeling box, improve labeling efficiency and reduce labeling difficulty, so that labeling personnel can complete continuous big data labeling of a specific scene in continuous time, improve data labeling accuracy and accelerate production of labeled data.

The application discloses a kind of quality evaluation and automatic labeling method and system of image data, comprising the following steps: S1: the original image of input is carried out multidimensional quality evaluation, and the image that does not satisfy multidimensional quality evaluation requirement is filtered according to preset threshold value;S2: the image screened out is carried out automatic labeling, and the preliminary segmentation mask is post-processed optimization by cross-validation iteration mechanism, make the segmentation result of prompt type segmentation model generated reverse correction positioning frame generated by target detection model, then utilize the positioning frame corrected after guiding prompt type segmentation model to carry out higher quality segmentation, output optimized automatic labeling result;S3: the automatic labeling result after optimization is refined and smoothed, and the standardized automatic labeling result is output;S4: the standardized labeling result and quality evaluation information are pushed to man-machine verification interactive system, receive artificial review or correction, and output final labeling file.

The application provides an image processing method and device, a management system, electronic equipment and a storage medium. The method comprises the following steps: obtaining a to-be-labeled image of a to-be-labeled region; obtaining labeling information of the to-be-labeled image based on a pre-trained labeling model, wherein the labeling model is obtained by pre-training a sample set with the labeling information; determining a target to-be-labeled image whose labeling information does not meet a preset requirement; and correcting the labeling information of the target to-be-labeled image based on the obtained correction information to meet the preset requirement. In this way, the labeling model can be used to label the to-be-labeled image, and the labeling information can be corrected based on this, which can reduce the labeling workload of the workers, improve the labeling efficiency, and ensure the recall rate and accuracy of the final labeling information.

The application discloses a kind of dense lesion semi-automatic labeling method for fundus image, method includes constructing and training dense lesion segmentation network, macula fovea positioning network and optic cup optic disc segmentation positioning network;Utilize network to process to obtain dense lesion prediction contour area, macular area and optic cup optic disc area respectively;Three labelers are assigned to fundus image, each labeler selects key area and other area from macular area and optic cup optic disc area and carries out labeling operation, to obtain first round labeling result;According to first round labeling result, the lesion labeling consistency index between every two labelers is calculated to determine two labelers to execute second round lesion labeling, and second round labeling result is obtained;Second round labeling result is audited to obtain final labeling result.Therefore, the problem of large difficulty in dense lesion labeling and large difference in labeling by different personnel is solved, and the labeling efficiency is improved, the labeling time is reduced, and the labeling cost is reduced.

The invention provides a multi-format label generation method and system and a storage medium, and the method comprises the steps: receiving label configuration information inputted through a visual interface, and analyzing the label configuration information to generate a corresponding declarative rule description language; arranging and executing a dynamic data processing flow based on the declarative rule description language to obtain dynamic data from a plurality of data sources; when the dynamic data processing flow is executed, caching operation is carried out on the component instances used in the dynamic data processing flow according to a predefined caching strategy to obtain reusable caching instances; the dynamic data and the cache instance are fused, and an internal document model is generated through precompilation instruction sequence rendering; and inputting the internal document model into a multi-format output pipeline, and converting the internal document model through a format filter to output final tag files in at least two different formats. The method and the device can flexibly and efficiently output various formats of files with consistent styles.

The invention discloses a printing method and device for generating a label picture based on configuration, a terminal and a storage medium, and relates to the field of data processing and image generation, and the method comprises the steps: obtaining a label definition configuration file and to-be-filled data, processing data conflicts, determining target data, calculating an initial pixel layout, and detecting layout conflicts. If conflicts exist, adjusting and generating a final layout, and finally synthesizing and outputting a final label picture according to the final layout. The technical effects of effectively processing data conflicts and layout conflicts and automatically generating and outputting high-quality final label pictures are achieved.

The present application relates to the technical field of semiconductor manufacturing quality control, and specifically relates to a space-time coupling virtual metrology intelligent labeling method based on feature space geometry, comprising the following steps: S1: obtaining multi-dimensional sensor time series data of a semiconductor manufacturing device as a labeling sample set, and performing manifold value evaluation on the labeling sample set; S2: establishing an adaptive time weight model based on data complexity, establishing a dynamic space weight model based on local sparsity based on an adaptive kernel function, combining the adaptive time weight model, the dynamic space weight model and the manifold value evaluation, and defining a space-time coupling value score; S3: taking the space-time coupling value score as input, establishing a three-layer optimization architecture to realize globally optimal sample combination; S4: merging the selected samples of each cluster to form a final labeling set, performing quality evaluation on the final labeling set, and generating a standardized labeling output according to the quality evaluation result; and providing an accurate value basis for subsequent intelligent selection.

The application discloses a rule intelligent labeling generation method based on task semantic analysis, relates to the technical field of data labeling management, and specifically comprises the following steps: performing zero-sample rule cold start, comprehensively calculating confidence scores according to occurrence frequencies, semantic rationality scores and domain general prior knowledge, and selecting a seed rule set; acquiring the seed rule set, constructing a rule knowledge graph based on multi-dimensional semantic association relationships; acquiring a natural language sentence to be labeled, performing bidirectional refining on sentence semantics and rule semantics of the natural language sentence through a cross-attention mechanism, and calculating a matching score; on the rule knowledge graph, label information is propagated from the seed rule to a candidate rule through an attention mechanism; the propagated label and the matching label are compared, inconsistent rules are marked and preferentially submitted for verification; the rule knowledge graph is updated according to a verification result, and a final labeling rule set is iteratively generated.

The present disclosure provides systems and techniques for handling long text for a pre-trained language model. In some aspects, a computing device may receive, at a data processing system, a set of utterances for training or inferring with a named entity recognizer to assign a tag to each lexical segment in the set of utterances. The computing device may determine a length of each utterance in the group, and when the length of the utterance exceeds a predetermined threshold for the lexical segment: divide the utterance into a plurality of overlapping lexical segment chunks; assigning a label and a confidence score to each lexical element fragment in the word block; determining a final tag and an associated confidence score for each lexical element segment word block by merging the two confidence scores; determining a final annotated tag of the utterance based at least on merging the two confidence scores; and storing the final annotated tag in a memory.

The invention discloses a consistency labeling decision-making method based on multiple algorithms. The method comprises the following steps: acquiring an image acquired by a vehicle-mounted camera; inputting the image into three annotation algorithm models to obtain an output result including an annotation bounding box, an annotation category and annotation confidence; matching the same object detected by different labeling algorithm models by taking the intersection-to-union ratio IoU of the labeling bounding box as a matching rule; calculating the similarity between output results of every two labeling algorithm models, and obtaining a consensus weight by using an average value of the similarities with other labeling algorithm models; and weighting object by object based on the consensus weight to obtain a final labeling result. According to the method, spatial consistency and confidence of recognition results of different labeling algorithms on specific samples are dynamically evaluated, and differentiated decision weights are intelligently given to the recognition results, so that serious divergence between models is effectively recognized and processed, uncertain labels are automatically screened out, and labeling errors caused by algorithm blind spots or accidental errors are reduced.

The invention provides a multi-modal data labeling method and device and a storage medium, and relates to the field of automatic driving, and the method comprises the steps: synchronously collecting multi-modal data, and obtaining environment parameters; based on the environmental parameters, obtaining the reliability of each modal, judging the modal with the highest modal reliability as a reliable modal, and judging other modals as unreliable modals; the pre-labeling result of the reliable mode is a final labeling result of the reliable mode; transforming the pre-labeling result of the reliable mode to obtain a projection result of the unreliable mode, wherein the projection result of the pre-labeling result of the reliable mode is projected on the unreliable mode; inputting the projection result of the unreliable mode and the pre-labeling result of the unreliable mode into a target association module, and outputting an intersection-to-union ratio; and correcting the pre-labeling result of the unreliable mode according to the intersection-parallel comparison to obtain a final labeling result of the unreliable mode. The method effectively solves the problems of poor environmental adaptability and insufficient multi-modal data fusion precision in the prior art.

The embodiment of the invention discloses a data annotation method and device. The method comprises the steps that firstly, any target sample in a to-be-labeled first sample set is processed through n first large models, corresponding n labeling results are obtained, and each labeling result comprises a labeled category label and a corresponding labeling basis; and then, processing the n labeling results by using a second large model to obtain a sorting result of m candidate labels involved in the n labeling results. Then, m confidence coefficients corresponding to the m candidate tags are determined, and any ith confidence coefficient is positively correlated with the sorting weight of the ith candidate tag determined based on the sorting result and negatively correlated with the confusion degree when the second big model generates the ith candidate tag; and then, under the condition that the maximum value in the m confidence coefficients is greater than a preset threshold value, determining the candidate label corresponding to the maximum value as the final label of the target sample.