790 results about "Task learning" patented technology

Provided in the present application is an intelligent question answering method based on a collaboration between a large language model and a knowledge graph, relating to the technical fields of artificial intelligence and natural language processing, the method comprising: decomposing a complex question into a plurality of simple questions, and analyzing the degree of association between the simple questions and a basic function so as to form a multi-hop reasoning path; automatically extracting structured information from the simple questions on the basis of a multi-task learning framework of a large model, so as to construct a knowledge graph; and constructing a cumulative reasoning learning framework on the basis of a logic reasoning large model, and performing iterative verification on a process result formed by the knowledge graph on the basis of the multi-hop reasoning path, so as to correct the reasoning path until a correct answer is inferred.

The invention belongs to the technical field of medical image processing, and discloses a CT image segmentation and classification system based on segmentation feature guidance, and the specific technical scheme is as follows: the system adopts a shared encoder to extract general features, realizes collaborative optimization of segmentation and classification through a double decoding path, adopts a partial decoder in a segmentation path, and adopts a partial decoder in the segmentation path; in combination with a local feature attention module, through multi-scale feature fusion and a boundary perception mechanism, the region consistency of a global segmentation map is gradually optimized, edge detail information is supplemented, and a classification path generates a space attention weight through a segmentation feature guide module by utilizing segmentation prediction; the classification network is guided to focus on a focus area and suppress background interference, a self-adaptive loss weighting strategy based on multi-task learning is adopted, the double-task gradient flow is dynamically balanced, and the gradient competition problem in the multi-task learning is effectively relieved.

The invention relates to the technical field of robot training. The robot self-adaptive training method based on reinforcement learning comprises the steps that task sub-target information is generated through a high-level strategy network, the task sub-target information is input into a low-level execution network, an action control instruction is generated according to the task sub-target information, interaction feedback information is collected in the execution process, and the action control instruction is sent to a robot through a robot. Calculating a reward value according to the interaction feedback information, carrying out association processing on the reward value and the scene complexity parameter, executing a dynamic reward shaping operation, generating an adjusted reward signal, generating a strategy model optimized by meta-learning based on the adjusted reward signal, loading the strategy model in a simulation environment, and carrying out dynamic reward shaping. A target strategy model optimized through simulation training is generated, the target strategy model is loaded to the robot, and the robot is controlled to execute task operation in the actual interaction scene. The method has the effect of realizing adaptive task learning of the robot in a multi-interaction scene.

The invention discloses a power equipment fault prediction system based on big data analysis. The method comprises the following steps: acquiring initial equipment multi-dimensional data; constructing a dynamic topology network of the power equipment, including a dependency relationship between the equipment and a fault propagation path, performing embedded learning on the dynamic topology network by using a GNN graph neural network, and extracting equipment collaboration features in the initial equipment multi-dimensional data; a multi-task learning framework is constructed in combination with the equipment cooperation features to predict the equipment fault probability and the remaining service life, and an equipment health index is obtained; and acquiring environmental parameters, dynamically adjusting a fault judgment threshold based on the equipment health index and the environmental parameters, generating a prediction result, integrating the prediction result with an SCADA system, and triggering graded early warning. And the influence of environmental factors on the operation state of the equipment is fully considered. Under different environmental conditions, the equipment fault risk can be judged timely and accurately.

The invention provides a city calculation basic model and equipment based on a Mama time sequence, and relates to the technical field of data processing. In the model provided by the embodiment of the invention, a Mama framework is combined with multi-task time sequence analysis. Aiming at the non-stationarity of urban time series data, a frequency prompt network is designed for decoupling in combination with a Koopman operator theory and dynamic mode decomposition, a steady mode and non-stationary fluctuation are analyzed from the angles of time and frequency, and a frequency mode memory pool (FPMP) is introduced for realizing cross-scene feature reuse. The advantage of a Mama model in capturing long-term time dependence is utilized, and an attention mechanism similar to Mama is designed to simulate channel dependence in a multivariable time sequence. A learnable supplementary sequence and a multi-task prompt mark are introduced to support dynamic multi-task adaptation, and the contradiction between urban data isomerism and task diversity is relieved. Shared knowledge and specific task knowledge are balanced through shared experts and specific task experts of the multi-task attention module, the negative migration problem in multi-task learning is relieved, and the generalization ability of the model is enhanced.

The invention discloses a farmland field segmentation method and system based on multispectral SAM model and multitask learning guidance, and belongs to the field of field segmentation. The problem that in the prior art, channel number matching is conducted in a simple channel compression or deep convolution mode, and the challenges that the boundaries of cultivated land parcels are complex, diversified and subtle are difficult to effectively deal with is solved. The method comprises the following steps: performing spectrum compression on a multispectral remote sensing image through a U-Net structure fused with a CBAM attention mechanism to generate a three-channel pseudo visible light image; the image is input into an SAM-ViT encoder subjected to LoRA fine tuning, and global semantic features are extracted; a task token sequence is constructed, explicit interaction between the token and the image features is realized through a bidirectional Transform module, and high-resolution fusion features are generated; calculating cross-task attention, and outputting a multi-task mask through inner product operation; and performing post-processing by using a watershed algorithm to generate a field boundary. The method is used in the field of cultivated land identification.

The invention relates to the field of artificial intelligence and big data processing, and discloses a professional field agent construction method and system based on a multi-modal hybrid expert model and a knowledge graph. The method comprises the following steps: collecting multi-source heterogeneous data such as user behaviors, texts, images, structured data and voices, and vectorizing and coding the multi-source heterogeneous data; performing feature extraction on different modal data and fusing to form unified representation; constructing a hybrid expert network based on a gating mechanism to perform task specialization processing on the fusion representation; introducing a domain knowledge graph and performing semantic enhancement through a graph neural network; on the basis of fusion representation, tasks such as recommendation, intention recognition and question and answer are executed at the same time through a multi-task learning framework; and an output result is fed back through the multi-modal interaction interface, and user feedback is collected to update the model. According to the method, the semantic understanding, task prediction and personalized recommendation capabilities of the model in a complex scene can be improved.

The invention discloses a sewage treatment plant effluent prediction method based on multi-task learning. The method comprises the following steps: acquiring sewage treatment data; based on the sewage treatment data, establishing an effluent prediction model; the input of the effluent prediction model is inflow water quality data, process data, environmental data and sewage treatment unit data, and the output of the effluent prediction model is predicted effluent index data; predicting the water outlet index data in future time based on the water outlet prediction model and the input of the water outlet prediction model; according to the method, the water outlet prediction model is constructed in combination with multi-task learning, and the model can comprehensively consider the time sequence dependence and mutual influence relationship among the input data to perform prediction, so that the calculation redundancy is reduced, and the prediction demand that an actual process needs to cooperatively consider multiple targets is met; the effluent quality prediction precision and the process regulation and control efficiency are remarkably improved, and a solid foundation is laid for promoting intelligence of operation management of a sewage treatment plant.

The invention discloses a satellite orbit forecasting method based on deep learning physical constraint loss, and the method comprises the following steps: 1, carrying out the normalization preprocessing of input data, forming a training data set and a test data set, and constructing batch processing training data; and 2, performing dimension expansion on sample data points in each window in the batch processing data formed in the step 1, constructing a multi-dimensional feature space of the sample points, and forming a batch processing input data format capable of being introduced into the model. And 3, performing forward reasoning on the batch data formed in the step 2 by using a model, and obtaining a batch processing orbit prediction value output by the model at the next moment through a CNN lightweight spatial-temporal feature extraction module and a BiLSTM bidirectional time sequence neural network module. And 4, taking the track prediction value obtained in the step 3 and the truth value label in the training set obtained in the step 1 as input, calculating to obtain a loss value of a current training iteration batch through a multi-random learning loss module fusing physical constraints, and performing reverse updating of model parameters to complete model training. And step five, through the steps two to four, performing reasoning verification on the model by using the test set formed in the step one, and comparing with a truth value in the test set to obtain a model test result.

Techniques are disclosed for training a model, using multi-task learning, to transform natural language to a logical form. In one particular aspect, a method includes accessing a first set of utterances that have non-follow-up utterances and a second set of utterances that have initial utterances and associated one or more follow-up utterances and training a model for translating an utterance to a logical form. The training is a joint training process that includes calculating a first loss for a first semantic parsing task based on one or more non-follow-up utterances from the first set of utterances, calculating a second loss for a second semantic parsing task based on one or more initial utterances and associated one or more follow-up utterances from the second set of utterances, combining the first and second losses to obtain a final loss, and updating model parameters of the model based on the final loss.

The invention relates to the technical field of psychological health, and discloses a psychological health multi-modal evaluation method, system and device and computer equipment, and the method comprises the steps: obtaining multi-modal data, and carrying out the feature extraction of the multi-modal data, and obtaining multi-modal features; performing emotion-oriented cross-modal attention mechanism analysis on the multi-modal features to respectively obtain a preset-dimension emotion scale vector, an emotion semantic feature, a multi-dimensional emotion feature and an uncertainty quantitative feature; fusing the preset dimension emotion scale vector, the emotion semantic feature, the multi-dimensional emotion feature and the uncertainty quantitative feature to obtain a multi-modal fusion feature; and obtaining a mental health assessment result based on the multi-modal fusion features and a preset multi-task learning framework. Through a cross-modal attention mechanism, deep semantic fusion of five modals of vision, audio, physiology, text and behavior is realized, and intelligent mapping from original multi-modal data to accurate psychological state judgment is realized through a deep learning technology.

The invention provides a software vulnerability understanding method based on adversarial training data enhancement and multi-task learning, belongs to the technical field of software quality assurance, and solves the technical problems that vulnerability type prediction and row-level vulnerability positioning are difficult to accurately complete at the same time and code semantics are easy to damage due to sample enhancement in the existing method. According to the technical scheme, the method comprises the following steps: (1) constructing a C / C + + vulnerability data set containing vulnerability type tags and row-level vulnerability tags; (2) generating an adversarial training data enhancement method constrained by AST and PDG in the embedding space by adopting a PGD algorithm; and (3) taking the pre-training model as a shared encoder. The method has the advantages that the performance of the two tasks can be remarkably improved, and the manual workload needed in the vulnerability understanding and analysis process is effectively reduced.

The invention discloses a spine image key point detection algorithm based on multi-task learning. The algorithm comprises the following steps: constructing a multi-task deep learning network model comprising a segmentation branch and a key point positioning branch; a feature aggregation module based on multi-scale cavity convolution is embedded in the segmented branches, and the multi-scale cone feature extraction capability of the model is enhanced through splicing fusion of multiple cavity rate convolution branches and global pooling branches; a cross-task attention fusion module is introduced between the two branches, and bidirectional dynamic interaction and complementation between segmentation features and key point features are realized by generating and fusing first-order and second-order context attention maps; semantic alignment loss is designed in a training stage, collaborative optimization of two tasks is promoted by constraining the consistency of segmentation masks and key point heat maps in a high-level feature space, global information of segmentation and local information of key point detection are fully utilized, and the accuracy and stability of spine centrum key point positioning and the accuracy of a segmentation result are improved.

The invention particularly relates to a multi-task motion recognition method based on a physical constraint guide condition diffusion model, which comprises the following steps: acquiring original observation data of multi-task motion based on an inertial measurement unit and preprocessing to obtain preprocessed original observation data; constructing a conditional diffusion model, and training the conditional diffusion model by using the preprocessed original observation data based on a preset loss function to obtain a trained conditional diffusion model; performing data enhancement on a key motion state category determined based on the motion state category label corresponding to the original observation data according to the trained conditional diffusion model to obtain a key motion state category sample, and constructing a sample balance data set based on the key motion state category sample; and based on the sample balance data set and a preset multi-task learning network loss function, constructing a multi-task learning pedestrian motion recognition model and carrying out multi-task recognition. Therefore, the problems of few key state samples, unreasonable data and the like in motion recognition are solved.

The invention relates to the technical field of medical image segmentation, and particularly discloses a pathological myopia analysis method and system based on multi-task learning, and the method comprises the steps: inputting a pathological myopia eye bottom image into a shared feature encoder for feature extraction, and obtaining an image embedding vector; the shared feature encoder is based on a Swin Transform architecture, and the rank value of the LoRA parameter is dynamically adjusted through the information entropy; performing feature enhancement on the image embedding vector through a layered progressive head attention module to obtain an enhanced image embedding vector; and inputting the enhanced image embedding vector into a multi-task solution terminal to obtain an analysis result. According to the method, the diagnosis precision and the calculation efficiency can be remarkably improved, powerful technical support is provided for automatic diagnosis and personalized treatment of ophthalmic diseases, and wide application of the intelligent medical technology in clinical practice is promoted.

The invention discloses an intelligent reasoning knowledge base construction method based on a knowledge graph, and relates to the technical field of government affair information processing and intelligent reasoning. Accurate extraction of strategy elements and formalized expression of logic relations are realized by constructing a structured strategy clause knowledge graph; a reliable data basis is provided for conflict detection; a method of combining a self-adaptive threshold mechanism and constraint satisfiability solution is adopted, so that the accuracy of conflict detection is ensured, and the method can adapt to the language characteristics and time evolution characteristics of a strategy text; clear decision support is provided for a strategy making department by establishing a complete conflict report generation mechanism; the timeliness and the consistency of the knowledge base are ensured by an incremental graph updating and version management function, and the department cooperation capability is further enhanced by the construction of the government affair responsibility sub-graph; the application of the multi-task learning model improves the intellectualization level of demand processing, and the personalized PageRank algorithm realizes accurate department recommendation.

The invention discloses a signal feature compression method based on codebook discrete quantization and multi-task learning, and belongs to the technical field of crossing of signal processing and artificial intelligence. In order to solve the problem that in the prior art, compression efficiency, semantic retention and calculation complexity are difficult to consider in high-dimensional IQ signal compression, high-dimensional continuous features of an original IQ signal are extracted through an encoder; carrying out vector quantization by utilizing the learnable codebook to generate a discrete index and calculating quantization loss; inputting the discrete features into a decoder branch reconstruction signal to calculate reconstruction loss, and inputting a classifier branch prediction category to calculate classification loss; processing traditional features and coding features by adopting a multi-layer perceptron, and calculating comparison loss through cosine similarity; combining optimization quantization loss, reconstruction loss, classification loss and comparison loss to train a model; and finally outputting a discrete index as a compression feature. The method realizes efficient compression and classification semantic reservation, and is suitable for wireless communication, Internet of Things and other scenes.

The invention discloses an electric power system transient stability state and stability margin evaluation method, device and equipment based on Grapher multitask learning and a medium, and relates to the technical field of electric power, and the method comprises the steps: carrying out the standardization of node features corresponding to a bus and edge features related to a power transmission line and a transformer branch, and obtaining standardized features; determining a dichotomy label based on a power angle difference between the generators after a fault occurs, determining a regression label based on a difference between a generator rotor angle trajectory and an inertia center trajectory, and determining a target Grapher multi-task learning model by using the standardized features, the dichotomy label, the regression label and an initial Grapher multi-task learning model in combination with a multi-task loss function; and determining a transient stability state and a transient stability margin of the power system based on the real-time operation data and by using the target Grapher multi-task learning model. Therefore, the accuracy of transient stability evaluation is improved.

The invention discloses a small sample aerodynamic modeling method based on multi-task learning. Comprising the following steps: constructing a multi-task prediction model, integrating an auxiliary task network and a target task network, and enhancing the feature extraction capability of an encoder through an SE layer and an attention layer; acquiring multi-source aerodynamic force data and carrying out standardized preprocessing on the multi-source aerodynamic force data; a dynamic weight mechanism is designed, the influence of auxiliary task prediction on target task output is adaptively adjusted according to input features, and effective fusion of low-fidelity data and high-fidelity data is achieved; two-stage training is carried out to ensure the efficiency and stability of multi-task learning; the prediction capability of the model under the limited sample condition is verified; through the synergistic effect of multi-task knowledge migration, feature enhancement and dynamic task integration, in combination with the feature optimization characteristics of an SE layer and an attention mechanism, high-precision aerodynamic prediction is realized under the condition of limited samples, and an efficient and accurate aerodynamic prediction method is provided for aircraft design.

The invention relates to the technical field of target behavior prediction systems based on causal inference and multi-task learning, and particularly discloses a target behavior prediction system based on causal inference and multi-task learning. The system comprises a central coordination server and a plurality of participant clients, constructs a global causal graph through a federated causal discovery algorithm in a collaborative manner, determines a causal feature subset of each prediction task, and carries out multi-task model training through a federated average algorithm under the constraint. In the process, differential privacy and homomorphic encryption technologies are comprehensively applied to protect data privacy. According to the method, more accurate causal discovery and more reliable prediction model training can be realized on the premise of protecting data privacy of all parties.

The invention relates to the technical field of vehicle intelligent management, and discloses a vehicle state monitoring method and system based on deep learning, and the method comprises the steps: analyzing the physical dependence relation among vibration abnormality detection, loosening positioning and fatigue evaluation monitoring tasks based on structural mechanics constraints, and carrying out the monitoring of the vibration abnormality detection, the loosening positioning and the fatigue evaluation; generating a task dependence directed graph; inputting the multi-modal monitoring data matrix into a shared encoder of a multi-task learning network to generate a unified representation vector; respectively inputting the unified representation vector into a vibration anomaly detection decoder, a loose positioning decoder and a fatigue evaluation decoder, and outputting an initial prediction result of each task; and establishing a conditional probability model between task outputs based on the constraint relationship defined by the task dependent directed graph, and generating a consistency diagnosis result meeting physical constraints through maximum posteriori estimation. The technical problems of one-sided diagnosis results, mutual contradiction and low calculation efficiency are solved.

The invention discloses a multi-modal sensing fusion dam body structure state intelligent diagnosis and early warning system and method and application. The system comprises a multi-modal sensing network composed of a space deformation monitoring subsystem, an internal response monitoring subsystem and an environment quantity monitoring subsystem; a data fusion and feature extraction module based on a space-time diagram attention network, which is used for deeply mining a complex space-time coupling relationship between multiple measurement points and multiple physical quantities; the health state diagnosis module based on multi-task learning can synchronously output dam body structure health indexes and abnormal types of key areas; and a future state evolution prediction and multi-stage early warning module based on a Transform model. According to the method, digital twinning of the dam body is constructed, a physical entity and an information model are closely combined, a full-chain closed loop from multi-dimensional perception to intelligent diagnosis to prospective early warning is achieved, and the accuracy, comprehensiveness and timeliness of dam body structure state evaluation are improved.

The invention discloses a ground and satellite observation constraint combined GPP multi-task learning estimation method, system, medium and equipment, and belongs to the technical field of remote sensing, the method comprises the steps of obtaining OCO-2 SIF data and TROPOMI SIF data, outputting a global GPP product prediction value through an expert network, and obtaining a global GPP product prediction result; the processing flow comprises the following steps: extracting initial features based on an OCO-2 SIF data set and a TROPOMI SIF data set, inputting the initial features into a shared encoder module, learning cross-sensor spatial and temporal features through a feature alignment and fusion mechanism, outputting OCO-2 SIF time features and TROPOMI SIF spatial features, carrying out independent decoding and outputting predicted values, and outputting the predicted values. According to the global GPP high-precision estimation method, the problem of an error mode of high-value underestimation and low-value overestimation caused by insufficient space-time generalization due to single task model estimation is solved.

The invention is suitable for the technical field of industrial process fault diagnosis, and relates to a hot rolling process composite fault tracing method under a variable working condition, which comprises the following steps: S10, fusing the generality and personality characteristics of a composite fault under the variable working condition by adopting multi-task learning and deep learning; s20, on the basis of extraction and fusion of generalization and personality characteristics of the composite fault under variable working conditions, adopting transfer learning and a convolutional neural network to realize intelligent diagnosis of the composite fault of working condition transfer and generalization; and S30, adopting an ensemble learning and transfer learning method to realize compound fault accurate tracing under variable working conditions. The method is simple in process and convenient to operate, and the accuracy of compound fault tracing in the hot rolling process under the variable working conditions is effectively improved.

The invention discloses a meta-class increment transfer learning driven cross-domain lifelong intelligent diagnosis method and device and a medium, and relates to the field of intelligent diagnos.The method comprises the steps that a sensor vibration signal is adopted to construct an increment task, and support and query sets of a source domain and a target domain are divided; for an initial task, adopting an internal and external loop training strategy, training a meta-class increment transfer learning model based on a source domain support set, and adopting a target domain support set to adjust training parameters; for other incremental tasks, continuously carrying out incremental task learning by adopting a multi-stage fusion knowledge distillation strategy and an internal and external loop training strategy, training a meta-class incremental transfer learning model to obtain a model of incremental task training parameters, and adjusting the training parameters by adopting a target domain support set to obtain an incremental task training parameter model; and inputting the information to enhance the target domain query set to obtain a final class increment migration diagnosis result. The method can solve the problems that an existing class increment learning method is limited in diagnosis performance in a class increment migration scene, is easily influenced by disastrous forgetting and is insufficient in model generalization ability.

The invention discloses a multi-task learning-based retinal vessel segmentation and analysis method and system, and relates to the technical field of medical image processing, and the method comprises the steps: introducing an EMCAD module into an R-Unet model, and carrying out the retinal vessel segmentation; designing an area percentage quantization function; a multi-task learning framework RetiVas-MTNet is constructed and is used for auxiliary diagnosis of eye diseases. The EMCAD module fuses a channel attention CAB mechanism and a space attention SAB mechanism: 1) generating a channel weight through adaptive maximum pooling and average pooling, and strengthening key channel features; (2) context connection of local features is established through large kernel convolution, and the space attention weight is calculated; according to the invention, the EMCAD module is introduced, so that the segmentation precision of the microvessels is improved; the blood vessel area percentage is quantitatively designed, so that the crossing from qualitative segmentation to quantitative diagnosis is realized; a multi-task learning framework RetiVas-MTNet provides a comprehensive solution for automatic diagnosis of eye diseases.

The invention provides a neural network model co-evolution method based on model modularization and model merging, and belongs to the field of artificial intelligence. Comprising the following steps: (1) carrying out model modular decomposition based on gradient search; and through common optimization of a weight retention rate index and a performance index on a field pre-training set, extracting a function-related sparse module from the pre-training model. And (2) the module only updates the specific weight on the downstream task, so that knowledge updating is carried out on the specific field. And (3) knowledge fusion based on model merging. A sparse task vector is directly obtained by subtracting the weights of the fine-tuned modules from the weights of the pre-training models, and then a sparse weight updating matrix among the multiple modules is added back to the pre-training weights, so that a multi-task global model is obtained. According to the method, the mapping relation between neural network parameters and functions is defined, the model co-evolution effect is improved, and parameter conflicts in multi-task learning are relieved.

Methods, systems, and computer-readable storage media directed to a machine learning (ML) model training system for training ML models by leveraging a large language model (LLM) for knowledge distillation to provide training data and using multi-task learning to train ML models using the training data.

The invention discloses a hybrid expert model structured pruning and acceleration method and system based on micro expert sorting, belongs to the technical field of large language models, and solves the problems that an existing MoE pruning method cannot consider the requirements of fine-grained pruning, reasoning acceleration and structural analysis generalization at the same time, coarse-grained expert-level pruning damages the performance of a model, and the reliability of the model is poor. And fine-grained compression lacks speed increase and lacks a unified micro-analysis method. The method comprises the following steps: splitting each expert network in a hybrid expert model into a plurality of micro experts, and modeling the plurality of micro experts, so that the micro experts in different expert networks have comparability; sorting all the micro experts according to the energy indexes of the micro experts by adopting a micro expert sorting algorithm; and processing the sorted micro-experts by adopting a pruning algorithm, selecting a core micro-expert for reservation, and directly deleting the rest of the micro-experts. The method is suitable for application scenes such as edge calculation and multi-task learning.

Disclosed in the present invention is a road-scene depth completion method based on the guidance of semantic information and a color image. The method comprises: designing a dual-branch network consisting of a color-image-guided branch and a semantic-guided branch, and introducing semantic information into the network, so as to perform road-scene depth completion. The idea of multi-task learning is used, and a backbone is shared with a depth map prediction layer in the color-image-guided branch, such that the semantic information can be obtained simply by adding a semantic segmentation layer, without the need to add an entire network. Moreover, semantic labels generated by the semantic segmentation layer are fed into the semantic-guided branch as input, such that parameters of the layer can also be adaptively adjusted during network training. By integrating high-precision semantic information and an RGB image, the present invention completes a sparse depth map provided by LiDAR, thereby improving the accuracy and real-time performance of scene understanding in applications such as digital twin, virtual reality, digital infrastructure and intelligent transportation.