1414 results about "Data labeling" patented technology

The invention discloses a data annotation method and system of a collaborative computing architecture based on quantum computing, and belongs to the field of data annotation. The method comprises the steps that S1, multi-modal data are input and preprocessed; s2, extracting features of each mode after preprocessing; s3, coding the features of each mode into a quantum state, and carrying out mode fusion; s4, performing label reasoning on the quantum state after modal fusion, and performing label constraint optimization by using a quantum approximate optimization algorithm; s5, based on a quantum Bayesian network or an approximate causal graph generation method, generating explanation according to a modal contribution causal path, and deducing marginal contribution of each modal to final label prediction by using a joint probability measurement result; and S6, outputting a labeling result. According to the method, a quantum-classical cooperative computing architecture is designed, the efficiency and accuracy of multi-modal data labeling are remarkably improved, the interpretability, the distributed processing capacity and the high-dimensional feature modeling capacity of the system are enhanced, and a brand new solution thought is provided for development of the multi-modal labeling technology.

The invention discloses a semantic fingerprint adaptive training method for a teaching service robot, and relates to the technical field of education neural network real-time training. Comprising six steps of course version semantic fingerprint injection, real-time drift detection and bucket division positioning, small sample correction and high-level weight patching, hierarchical control incremental learning scheduling, shadow reasoning consistency optimization and learning asset registration and cycle verification and tracking. Measuring drifting in real time by using an information entropy self-adaptive window and multi-scale divergence, generating a lightweight weight patch by small sample contrast learning, and performing online loading; shadow channel parallel reasoning is combined with a grading heat exchange superior weight, four-dimensional learning asset tensor is written into a registry through double-clock witness and chain commitment, random sampling verification and singular value performance verification ensure that assets are consistent with robot online examples, and the comprehensive effects of source traceability, risk self-sensing, model self-repairing and compliance full-chain trace reserving are achieved.

Provided is a detection method of spaceborne GNSS-R original intermediate frequency coherent reflection signals in ocean, polar and inland water areas, including: acquiring spaceborne GNSS-R original intermediate frequency signal data of TDS-1 or CYGNSS and preprocessing the data; selecting coherent detection feature engineering; setting data labels of different scenes and coherent and incoherent reflected signals; dividing a training set and a test set; and training and testing a multimode-oriented hybrid model for coherent and incoherent detection and classification of spaceborne GNSS-R signals, using the training set to train a model, applying a trained detection model to a test data set, and comparing and evaluating obtained detection results with a classical coherent detection algorithm.

The invention provides an intelligent data labeling method and system based on multi-modal fusion and large model verification, belongs to the field of artificial intelligence and data processing, and innovatively fuses multi-modal information such as an OCR recognition result, a layout structure, original image visual features and deep semantic analysis of a large language model (LLM). And a precise automatic labeling result credibility evaluation mechanism is constructed. According to the method, various errors in automatic labeling can be accurately recognized and adaptively corrected, and the errors comprise conventional error correction based on hard coding rules and complex semantic error correction driven by LLM. Meanwhile, the system can continuously optimize the data labeling capability of the system through an efficient man-machine cooperation and closed-loop feedback learning mechanism, and automatically precipitate domain knowledge assets. The invention aims to solve the problems of recognition accuracy bottleneck, heavy manual proofreading burden, lack of intelligent judgment and error correction, knowledge accumulation lag and the like in traditional document data labeling, so that the efficiency, accuracy and automation level of document data labeling are remarkably improved.

The invention provides an anomaly detection multi-classification method based on multi-source operation and maintenance data fusion. Comprising a data input layer, a parallel coding layer realized through dissimilatory multi-modal coding and a hierarchical multi-modal fusion architecture, a space-time feature fusion layer realized through a space-time perception dynamic gating attention enhancement mechanism, and a dynamic decision optimization layer realized through a gradient perception dynamic smooth loss function. The spatio-temporal feature fusion generates a feature representation and weight matrix with a dynamic attention weight through a spatio-temporal perception dynamic gating attention enhancement mechanism, and outputs the feature representation and weight matrix to the dynamic decision optimization layer; and the dynamic decision optimization layer realizes anomaly detection through a classifier taking a gradient perception dynamic smooth loss function as feedback, so that key problems such as multi-source heterogeneous data fusion, time sequence dynamic modeling and data label imbalance are solved, the anomaly detection accuracy and robustness of a training cluster are effectively improved, and the anomaly detection accuracy and robustness of the training cluster are improved. And a reliable technical support is provided for intelligent operation and maintenance of a complex training cluster.

The invention discloses a multi-modal data labeling method and system based on large model pre-labeling, and the method comprises the steps: S1, receiving to-be-labeled multi-modal original data and labeling task definition, and generating a structured task instruction signal; s2, inputting the structured task instruction signal into a multi-modal large model, and generating a pre-labeling result signal containing a preliminary label and a corresponding confidence coefficient thereof; s3, scheduling a manual verification task based on the confidence coefficient in the pre-labeled result signal; s4, according to the manual verification signal, performing parameter fine tuning or prompt optimization on the multi-modal large model, and generating a model optimization signal; and S5, pre-labeling the new multi-modal original data by using the multi-modal large model updated by the model optimization signal, and fusing an artificial verification signal. According to the multi-modal data annotation method and system based on large model pre-annotation, the problems that traditional multi-modal data annotation is low in efficiency, high in cost and difficult to unify in quality can be solved.

The invention belongs to the technical field of artificial intelligence, and discloses a reinforcement learning method for improving mathematical ability of a large language model and a related device. The reinforcement learning method comprises the steps of obtaining a to-be-enhanced large language model and a reinforcement learning data set; performing fine tuning training on the to-be-enhanced large language model by adopting reinforcement learning, and performing process level labeling on answer prediction generated in reinforcement learning by applying Monte Carlo estimation during fine tuning training to obtain a fine-tuned large language model and a labeled data set; and training the process reward model based on the annotation data set to obtain a trained process reward model. According to the technical scheme disclosed by the invention, fine-grained errors existing in the reasoning process can be captured, and the mathematical ability of a large language model is enhanced; in addition, data annotation can be realized while reinforcement learning is carried out, and the collection cost of process-level annotation data is saved.

The invention relates to the technical field of image recognition data processing, in particular to a pipeline element library creating method based on a model training and image recognition technology, which comprises the following steps of: acquiring pipeline element image data with depth information through a multispectral imaging scheme, executing Retinex algorithm illumination equalization and morphological restoration through a cascade preprocessing pipeline, and establishing a pipeline element library. A standardized data set associated with the metadata is generated. A multi-task joint learning framework is adopted to integrate ResNet-50, HRNet and Mask R-CNN networks, a bottom convolution feature extraction layer is shared, oil stain and strong reflection antagonism data synthesized by a generative adversarial network is injected, model parameters are optimized in combination with a progressive training strategy, and a lightweight MobileNetV3 model is output. Model parameters and an index structure are dynamically updated, and system self-calibration is achieved in combination with a cross-device calibration protocol and data consanguinity tracking. According to the method, through automatic data labeling, multi-task feature multiplexing and retrieval feedback closed loop, the construction efficiency of the pipeline element library is effectively improved, and the recognition error rate in a complex environment is reduced.

The invention provides a semiconductor chip automatic test optimization system, and belongs to the technical field of electrical variable measurement. Comprising an acquisition module used for establishing an electrical variable time sequence arranged according to a time sequence, a drift analysis module used for calculating a short-time fluctuation amplitude value, a long-term drift slope and a high-frequency noise amplitude value based on the impedance time sequence arranged according to the time sequence, and a contact judgment module. The evaluation module is used for constructing a contact impedance coefficient based on a short-time fluctuation amplitude value, a long-term drift slope and a high-frequency noise amplitude value in a jth sliding window, and evaluating and optimizing the contact impedance coefficient, and the data marking module is used for marking electrical variable measurement data corresponding to a test channel which is evaluated to be instable in contact as low confidence. According to the system, the ith test channel of the temperature sensor chip can be tested more accurately, the short-time fluctuation amplitude value, the long-term drift slope and the high-frequency noise amplitude value are calculated at the same time, multi-source data participates, and the test accuracy is improved.

The invention discloses an intelligent data labeling method based on artificial intelligence, and relates to the technical field of intelligent medical treatment, and the method comprises the following steps: obtaining physiological parameter data, historical health parameter data and motion state data of a user; outputting motion state label data through a pre-constructed motion recognition model according to the motion state data; and performing sliding window analysis on the historical health parameter data, and dynamically establishing and updating long-term change data of the user health parameters. According to the scheme, whether the physiological parameter fluctuation is caused by the movement or not can be distinguished through associated retrieval of time window division and the movement state label, for example, when it is detected that the heart rate fluctuation of the user during running exceeds the threshold value, normal physiological response instead of abnormity can be judged in combination with the movement label; the probability that normal physiological parameter fluctuation caused by movement is mistakenly marked as abnormal can be effectively reduced, and meanwhile, the pertinence of anomaly detection is improved.

The invention discloses a data labeling method and system based on user behaviors and attention tracking, and the method comprises the steps: synchronously collecting multi-source behavior signals of a mouse, a keyboard, eye movement and the like of a doctor in real time, combining identity and interface metadata, and carrying out the standardized normalization, abnormality elimination and short time sequence behavior unit division. And extracting individual behavior micro-modes by using unsupervised clustering, and constructing a behavior portrait library. Through multi-modal time sequence modeling and a self-adaptive space-time attention mechanism, behavior characteristics, an interface area and a report text are deeply fused, a multi-level correlation probability is output, and high-precision automatic tagging of content and an image area is realized.

The invention relates to the technical field of computer data processing, provides a cross-border data compliance processing method, device and equipment, and is used for solving the problems of relatively low cross-border data processing efficiency and data processing strategy errors in related technologies. According to the embodiment of the invention, the multi-method domain and regulation knowledge graph is established in advance, laws and regulations of a plurality of regions are summarized, the multi-method domain and regulation knowledge graph can be updated in real time, and the cross-border scene template library and the data label template base library are established according to the multi-method domain and regulation knowledge graph. The cross-border scene template library and the data label template base library can be dynamically updated along with updating of the knowledge graph, and a cross-border gateway can process data by adopting a compliance strategy according to a data cross-border scene through the cross-border scene template library, so that related data management enterprises and institutions efficiently manage data outbound in a compliance manner; through the data label template library, the cross-border gateway can rapidly and accurately carry out data marking work, and the efficiency of data classification processing is improved.

The invention discloses a multi-level dynamic isolation data processing system and method, and relates to the technical field of data security, the system comprises a data access module supporting a cloud native environment and a hybrid deployment mode, a strategy matching module, a sandbox execution module and an audit analysis module; the data access module receives the data, evaluates the sensitivity level and the purpose of the data, and labels the data based on an evaluation result; the strategy matching module reads the data label, searches an isolation strategy and an auditing rule which are matched with the data label from a strategy library, and automatically loads the strategies to the sandbox execution module once the matched strategies are found; the sandbox execution module selects an isolation mode according to the sensitivity level of the data, and processes and analyzes the data in an isolation environment according to a preset rule and algorithm; and the auditing analysis module records all operations on the data in real time, and analyzes the operation compliance based on a preset auditing rule. The problems of large resource consumption, insufficient strategy flexibility, coarse behavior audit granularity and the like in the existing data sandbox technology can be solved.

The invention discloses an information extraction data annotation and training system based on a large language model, and relates to the technical field of information extraction, and the system comprises a data set construction module which is used for constructing a pre-training data set and a fine tuning data set; the model continuous pre-training module is used for carrying out continuous pre-training on a preset general large language model based on the pre-training data set to generate a field adaptive pre-training model; the model fine tuning module is used for performing supervised fine tuning training on the domain adaptive pre-training model through a two-stage course learning strategy based on the fine tuning data set, and generating an information extraction model; the retrieval enhancement generation module is used for performing entity-semantic retrieval on an input text based on a preset knowledge base, outputting context information related to the input text, and outputting structured information of the input text based on the context information and an information extraction model, the problems of insufficient generalization ability, poor field adaptability and disastrous forgetting of a general large language model are solved, and the accuracy and robustness of information extraction are improved.

The invention discloses a data label generation and quality control method and system based on a pre-training large model, and relates to the technical field of data management. Aiming at the problems of low efficiency, poor label consistency and difficulty in continuously guaranteeing quality of an existing label generation mode, the adopted scheme comprises the following steps: collecting multi-source heterogeneous data based on a data governance demand; performing preprocessing and vectorization operation on the collected data, then outputting a pre-training large model, and realizing automatic generation of data labels by the pre-training large model in combination with rule constraints; performing quality detection of accuracy, consistency and coverage on the generated labels; performing optimization adjustment on the pre-trained large model or rule according to a quality detection result to realize feedback optimization of a label generation strategy; and outputting a label set passing the quality detection, and binding the label set with the original data to form a traceable label management record. The method is used for subsequent data management and large model training. The method is used for data governance and large model training.

The invention discloses a small model algorithm self-evolution optimization system and method based on multi-mode large model driving. According to the system and the method, a real-time image acquired by an automatic acquisition device is subjected to preliminary analysis, the acquired real-time image is subjected to visual enhancement processing, and image semantic analysis, cross-modal alignment and structured reasoning are performed to obtain image semantic features, text description and a spatial topological relation; then, multi-modal feature knowledge of the large model is migrated to the small model, feature level comparison is completed, and collaborative optimization parameters are synchronously generated; and finally, constructing a personalized training data set according to the collaborative optimization parameters, executing incremental model optimization based on transfer learning, and finally forming a detection-analysis-training closed loop iteration system. By using the system and the method of the invention to automatically label error samples, the workload of manual labeling can be reduced by about 60-80%, the data labeling cost is reduced, and the continuously updated data set can improve the adaptability of the model to new categories and complex scenes.

The invention discloses a structure characteristic quantitative extraction method and system for a titanium alloy microstructure image, and the method comprises the steps: obtaining the data of a titanium alloy microstructure grain image, and carrying out the data marking and image preprocessing; constructing and training a ResNet-based classification model, and distinguishing different tissues in various titanium alloy microstructure grain images; constructing a segmentation model TiGrainsU-Net by using the U-Net and an attention mechanism, inputting original microstructure grain image data and labeled mask data into the segmentation model for model training, and predicting different phase structures in the titanium alloy microstructure image by learning characteristics of different phase grains; performing multi-stage image optimization processing on a segmentation result, clustering different phases by using a final segmentation result, and calculating a quantization parameter; according to the method, the recognition capability of the model on boundary features and local detail features is improved, the segmentation accuracy is improved, and the precision and robustness of quantization parameter calculation are improved.

The invention discloses a big data-based finance and accounting data storage management optimization method and system, and relates to the technical field of financial information management, and the method comprises the steps: collecting enterprise financial original data, carrying out the multi-dimensional cleaning and structured preprocessing, recognizing the data type based on a classification model, and generating a metadata label; executing a cold and hot hierarchical storage and load balancing strategy in combination with the access heat and the node resource state; encryption, copying and compression operations are completed before data storage, and a multi-level index is constructed to improve the access efficiency; and finally realizing long-term archiving and system operation and maintenance management. The system comprises a data acquisition module, a preprocessing module, a classification module, a storage optimization module, a data storage module, a data access module and a system management module. According to the invention, intelligent management and efficient storage of massive heterogeneous financial and accounting data are realized, and the data processing capability, security and availability are improved.

The invention relates to the technical field of social data processing, in particular to a personalized recommendation-oriented big language model social data tag attribute generation method. Comprising the following steps: constructing a multi-modal social feature incidence matrix; generating an initial label set adapted to the social scene; and establishing a tag hierarchical relationship through an improved semantic dependency tree, calculating an aging attenuation coefficient based on a behavior occurrence time interval, and counting a proportion of users with the same tag in a core social circle as a group association value. According to the method, the complexity of the multi-modal data in the social scene can be adapted by constructing the multi-modal social feature incidence matrix, so that the generated tag system is more fit with the multivariate features of the social data. According to the method, the tag hierarchical relationship is automatically constructed through the improved semantic dependency tree, the aging attenuation coefficient is calculated, and the deep analysis capability of social semantics is improved.

The invention discloses an intelligent monitoring method for an annealing process of a wound transformer iron core, and particularly relates to the technical field of intelligent monitoring. The method comprises the following steps: acquiring temperature data and magnetic field intensity data generated in an annealing process of a wound transformer iron core, and carrying out time domain synchronization processing and abnormal data labeling; identifying an iron core micro-deformation area and a corresponding temperature anomaly area according to the synchronous data, and extracting temperature anomaly fluctuation characteristics; analyzing a dynamic coupling relationship between the temperature abnormal fluctuation characteristic and the iron core stress change, and outputting characteristic data of an iron core magnetic field and stress interaction; synchronously analyzing the interference intensity of abnormal temperature fluctuation on the annealing heat distribution uniformity, and outputting temperature disturbance intensity data; and constructing a coupling anomaly evolution model, generating iron core annealing anomaly evaluation data, quantifying a quality fluctuation risk in an annealing process, and generating intelligent early warning information. The monitoring precision and reliability of the iron core annealing process are effectively improved, and the annealing quality of the transformer iron core is guaranteed.

The invention provides a document element rapid extraction system based on a pre-trained large model, and relates to the technical field of computer software application, the system comprises a parameter field adaptation module used for textualizing a document and constructing an industry standard corpus based on a textualized processing result, adjusting a preset language model by utilizing an industrial standard corpus; the dynamic document partitioning module is used for performing semantic segmentation processing on the industrial standard document to obtain a plurality of text blocks; the entity alignment module is used for carrying out entity and relation extraction on the text blocks and carrying out entity alignment in combination with a uniform manifold approximation and projection method; and the relation reasoning and knowledge graph completion module is used for performing completion processing on the preliminary knowledge graph and storing a completion result. According to the method, the element extraction efficiency can be directly improved without pre-defining a rule template or performing data annotation.

The embodiment of the invention provides an annotation task allocation method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring a plurality of annotation tasks to be distributed; determining task portraits of the plurality of to-be-distributed annotation tasks according to the plurality of to-be-distributed annotation tasks; obtaining capability portraits of a plurality of annotation personnel and capability portraits of a plurality of data annotation models, wherein the data annotation models are used for data annotation; according to the task portrait, the ability portraits of the plurality of annotation personnel and the ability portraits of the plurality of data annotation models, constructing an optimization model used for determining an allocation relationship among the to-be-allocated annotation task, the annotation personnel and the data annotation models; and solving the optimization model by using a solving algorithm to generate a target task allocation scheme.

The invention discloses a personalized customization data mining and analysis system based on artificial intelligence, and the system comprises the following modules: an edge data collection module which is used for generating a structured multi-source heterogeneous data set; the high-order data tensor construction module is used for constructing a high-order data tensor based on the structured multi-source heterogeneous data set; the heterogeneous tensor decomposition module is used for extracting a potential feature matrix and a core tensor; the feature fusion and unified expression module is used for organizing all the fused feature vectors according to a time sequence to form a unified feature expression sequence; the dynamic clustering analysis module is used for generating a clustering evolution diagram; the clustering stability evaluation and feedback module is used for obtaining a fusion feature vector in a cluster structure mutation state; and the abnormal data labeling module outputs an abnormal data labeling result set. According to the method, the overall data availability and maintainability of the system are greatly improved.

The invention relates to the technical field of data processing, in particular to a multi-modal data labeling method, which comprises the following steps of: determining a data type and basic characteristics of data to be labeled to determine an initial association anchor point group; determining cross-modal semantic similarity according to keyword features, visual features and time sequence features in the initial association anchor point group, and correcting a keyword feature text in combination with visual feature confidence to update the initial association anchor point group; constructing a differentiated knowledge graph, and performing candidate anchor point association in combination with a preset semantic association rule to determine an expansion association confidence so as to determine an anchor point expansion processing mode; and carrying out statistics on the association group activation frequency of each initial association anchor group in a plurality of time sub-segments in a preset labeling period, determining the activeness of each initial association anchor group in combination with the time sub-segment attenuation weight, and determining a recommendation mode of each initial association anchor group according to the activeness and the association group activation frequency. According to the method, cross-modal semantic alignment is realized, and the accuracy of multi-modal data labeling is improved.

Technologies and techniques for converting sensor data, used in a vehicle or other device. A machine-learning model is applied to first sensor data, including a first operational characteristic capability and first sensor label data, wherein the machine-learning model is trained to second sensor data including a second operational characteristic capability. New sensor data is generated that corresponds to the applied machine-learning model, wherein the new sensor data includes translated first sensor label data. A loss function may be applied to the new sensor data to determine the accuracy of the new sensor data and translated first sensor label data. In some examples, a multi-dimensional matrix of camera sensor parameters may be applied to the first sensor data labels to transform the first sensor data labels to second sensor data labels.

The invention discloses a fault diagnosis method under an imperfect information condition based on orthogonal space heterogeneous mapping, and relates to the technical field of power system key component fault diagnosis. A hierarchical fusion multi-dimensional fault feature extraction network based on a cross domain attention mechanism is provided and is used for extracting features of a complete data set in a fault set, and a multi-dimensional feature vector encoder is formed. For incomplete and missing data, an enhanced orthogonal spatial heterogeneous network embedded with a multi-dimensional feature vector is provided, a data reconstruction process is optimized by introducing a multi-dimensional feature vector encoder and an identical mapping residual block and integrating a fault feature frequency and a generative adversarial reconstruction loss function, and the integrity of the data is enhanced. For missing and wrong labels, a label self-correction strategy based on space differentiation is provided, the strategy not only considers space information feature matching between different fault types and different sensors, but also provides space differentiation feature frequency projection, and therefore the accuracy of data labels is improved.

The invention provides a data marking method. The method can be applied to the technical fields of big data and artificial intelligence. The method comprises the steps of obtaining multiple pieces of multi-modal data, preprocessing the multiple pieces of multi-modal data, and generating multiple pieces of preprocessed text data; and performing vector conversion on the plurality of pieces of preprocessed text data to generate a plurality of pieces of feature vector data. Density clustering is carried out on the multiple pieces of feature vector data, a data sets are generated, and each data set comprises a first data label. Semantic clustering is conducted on the a first data labels through semantic analysis, and b second data labels are generated. And presetting a business knowledge graph, and performing knowledge fusion on the business knowledge graph and the b second data tags to generate b target data tags for data marking. The invention further provides a data marking device and equipment, a storage medium and a program product.

The invention discloses an automatic driving long-short time decision data labeling method and device based on a conflict area, electronic equipment and a storage medium, and the method comprises the steps: in response to own vehicle track data and other vehicle track data, dividing a decision scene to obtain a game type label, and calculating the conflict area information of an own vehicle and other vehicles; marking long and short time decision information according to the game type label and the conflict area information of the own vehicle and the other vehicle; and obtaining a group of decision label data according to the game type label, the conflict area information of the own vehicle and the other vehicle, and the long and short time decision information. According to the decision data labeling method provided by the invention, a richer game scene is provided, so that the labeling data not only labels corresponding decision behaviors, but also labels time margins and space margins of specific robbing and giving.

The invention provides a data labeling method and system, and the method comprises the steps: carrying out the voxelization processing of a laser radar point cloud and a three-dimensional image of a target scene, and obtaining multi-scale voxel features; performing sparse interaction on the point cloud voxel features and the image voxel features of the same scale to generate multi-modal sparse interaction voxel features; determining a three-dimensional bounding box of at least one target object in the target scene and a corresponding confidence coefficient; according to a natural language instruction provided by a user and used for describing a target object, generating a description text corresponding to the natural language instruction through the large language model; driving a preset prompt type visual model based on the description text, and generating a two-dimensional segmentation mask corresponding to the target object through an iterative interaction mode; according to the method, the three-dimensional bounding box is matched with the two-dimensional segmentation mask, and the semantic category represented by the two-dimensional segmentation mask is labeled to the matched three-dimensional bounding box, so that the manual intervention frequency is reduced, and the labeling efficiency, flexibility and consistency are improved.

The invention belongs to the technical field of data processing, and provides a data labeling method and system based on cue word driving, a joint guide vector is generated by fusing field features of a labeling task and operation behavior vectors of labeling personnel, labeling errors caused by guide deviation are greatly reduced, and labeling efficiency and preliminary labeling quality are improved; the initial labeling result is analyzed, the domain distribution difference between the labeling defect type and the associated cue word is recognized, and clear targeting is provided for follow-up knowledge graph parameter optimization; a reward function is constructed by taking a labeling defect type and cue word field distribution difference as a state space and combining labeling accuracy and field adaptability, parameters of a cue word knowledge graph are iteratively corrected through reinforcement learning, and high-quality cue words can be continuously output; after the cue word knowledge graph iteration is stable, the fusion coefficient of the joint guide vector is updated based on quality evaluation data feedback, and it is ensured that the joint guide vector and the optimized knowledge graph are cooperatively matched.