692 results about "Domain model" patented technology

The invention relates to a ship safety monitoring method, system and device based on data fusion, and relates to the technical field of ship navigation safety, and the ship safety monitoring method comprises the steps: collecting multi-source heterogeneous data; preprocessing the multi-source heterogeneous data; carrying out fusion processing on the multi-source heterogeneous data based on an information entropy theory; a ship safety domain model is constructed, trajectory prediction is executed, a collision risk index is calculated, and the ship collision risk is evaluated; and performing anomaly detection on the ship equipment operation data based on manifold learning and a thermodynamic model. According to the invention, a multi-source heterogeneous data fusion technology is adopted, multi-dimensional data sources are integrated, and omnibearing ship safety situation awareness is constructed. The early warning time is prolonged to 8-10 minutes, which is far better than the early warning time of 3-5 minutes in the prior art. The system thoroughly solves the problem that subsystems in a traditional ship monitoring system operate independently and do not communicate with one another, the crew information acquisition cost is reduced, and the number of times of interactive operation is reduced by 42%.

The invention discloses an adaptive medical image segmentation method during testing based on prototype alignment. The method comprises the following steps: acquiring target domain medical image data; obtaining a source domain model, wherein the source domain model comprises a feature extractor and a classifier; extracting a class prototype of each class by using a classifier, and initializing two models with the same structure as the source domain model by using a feature extractor and the classifier of the source domain model; the two initialization models are respectively used as a teacher model and a student model; the target domain image is input into a teacher model and a student model, the teacher model outputs a pseudo tag and an entropy graph of a target domain, the student model carries out supervised training according to the pseudo tag, and a target domain prototype output by the student model is aligned with a source domain prototype; the teacher model is updated through the index moving average value of the student model parameters, and finally the teacher model outputs a segmentation result. According to the method, the relation between the source domain and the target domain can be effectively established, and the cross-modal image segmentation precision is improved.

The invention discloses a passive unsupervised domain adaptive medical image segmentation method and device based on pseudo labels, and relates to the technical field of computers, and the method comprises a supervised source domain model training stage, an unsupervised adaptive target domain model training stage and a segmentation stage. The method specifically comprises the following steps of: performing pixel-level self-marking on a source domain medical image by applying pseudo-tag pixel-level uncertainty estimation to generate a reliable pseudo-tag; and on this basis, a strong and weak enhanced teacher-student target domain model is used, and a high-quality pseudo tag is fully utilized to train a target domain medical image, so that the network adapts to representation specific to the target domain medical image. According to the method, unsupervised segmentation is carried out on the target domain data graph on the premise that the source domain model and the target domain unmarked data are known, source domain data do not need to be accessed, the features of the source domain model can be reasonably utilized, and the problem existing in SFUDA can be well relieved; and the accurate segmentation capability on the target domain is still maintained under the condition that less data is available compared with the UDA method.

The invention provides a numerical simulation method suitable for a three-dimensional spherical flame, and belongs to the field of electrical data processing, and the method comprises the steps: building a computational domain model according to the requirements of a computational scene, dividing a grid unit, and obtaining the node information of the required grid unit; establishing a Navier-Stokes equation set of a coupling chemical reaction and diffusion process model; estimating convection and diffusion term variable gradients in the conservation equation; calculating a time step length according to a set Kulang number, and carrying out time advancing; estimating an expected value of a velocity field, and solving a predicted velocity; the divergence of the velocity field and the correction value of the pressure field are calculated; obtaining a corrected velocity field; a self-adaptive grid means is introduced to dynamically adjust the grid resolution; performing iterative correction for multiple times until the velocity field and the pressure field converge to convergence conditions; performing spherical flame numerical simulation, and monitoring the convergence of the solution. According to the invention, the flame dynamics and cellular instability of the premixed spherical expansion flame can be reproduced.

Disclosed in embodiments of the present description are a transfer learning model training method and device, a storage medium, and a terminal. The method comprises: at a target transfer moment, initializing a source domain model and a target domain model which satisfy a preset similarity condition; on the basis of parameter transfer networks corresponding to target domain hidden layers in the target domain model, obtaining target source domain hidden parameters in the source domain model; and transferring the target source domain hidden parameters to obtain a target domain transfer model, and training the target domain transfer model on the basis of target sample data corresponding to a target domain. Target source domain hidden parameters reflecting deep network features are selected from source domain hidden layers of the source domain model by means of the parameter transfer networks and transferred to corresponding target domain hidden layers of the target domain model, so that the target domain hidden layers directly obtain effective deep feature parameters, and the target domain transfer model is still trained by using the target sample data of the target domain.

The systems and methods disclosed herein relate to querying data using artificial intelligence models. A generalized model receives an output generation request and partitions it into segments mapped to specific domains, where each domain indicates associated databases and guidelines. The segments are routed to domain-specific models trained on domain-specific data, which generate query fragments by comparing performance metrics and system resource usage metrics. The query fragments are aggregated into an overall query that satisfies guidelines across domains. The systems and methods can include a feedback loop to adjust the domain-specific models using user interactions and performance metrics to dynamically adapt to a skill level or experience of the user.

The invention provides an intelligent discriminant optimization system and method based on a large language model, and the system comprises an intelligent agent management module which is used for managing various types of intelligent agents, and comprises an intelligent discriminant and a plurality of business intelligent agents; the big language model calling module is used for processing different business prompt cues according to the indication of the intelligent discriminant and interacting with corresponding big language models; and the large language model management module comprises three types of models, namely a small parameter model, a large parameter model and a vertical domain model. According to the method, the most suitable large language model is selected for processing, each sub-task is ensured to be processed by the most skilled model, and the accuracy of a processing result is improved.

The invention discloses a consultation method and system based on natural language processing and a legal knowledge graph, and relates to the field of data processing, and the method comprises the steps: receiving a multi-format legal consultation demand of a user, converting the multi-format legal consultation demand into a text, inputting the text into a BERT law NLP model, and analyzing key information through word segmentation, intention recognition and entity extraction; based on a pre-constructed multi-level legal knowledge graph, carrying out accurate and fuzzy retrieval and domain filtering in combination with an analysis result, and obtaining an associated law article, a case and a legal relationship; screening conflict law articles and similar cases, and inputting the conflict law articles and the similar cases into a graph neural network reasoning model to generate a preliminary conclusion; the conclusion is converted into a spoken consultation report through a natural language generation module, and output is customized according to a user scene; and if the user feedback satisfaction degree is less than the threshold value, iteratively optimizing the storage data to the historical library. The method has the advantages that accurate retrieval is realized based on the BERT model and the multi-level knowledge graph in the legal field, the oral personalized conclusion combined with the user scene is generated through GNN reasoning, and iterative optimization is performed through user feedback.

This specification is generally directed to techniques for generating interfacing source code between computing components based on natural language input. In various implementations, a natural language input that requests generation of interfacing source code to logically couple a first computing component with a second computing component may be processed to generate an interface request semantic embedding. The interface request semantic embedding may be processed based on one or more domain models associated with the first and second computing components to generate a pool(s) of candidate code snippets for logically coupling with first and second computing components. A plurality of candidate instances of interfacing source code may be generated between the first and second computing components. Each candidate software interface may include a different permutation of candidate code snippets from the pool(s) of candidate code snippets.

The invention provides a large model training method and device based on privacy protection and a storage medium. The method comprises the steps of splitting a large model into a plaintext domain model and a secret state domain model; receiving a private data ciphertext sent by at least one client; the private key ModelSK is utilized to decrypt the private data ciphertext in the TE to obtain private data; training the secret state domain model and the plaintext domain model by using the privacy data to obtain the trained secret state domain model and plaintext domain model; and encrypting the trained secret state domain model based on the determined client public key UserPK, and sending the trained plaintext domain model and the encrypted secret state domain model to the client. Therefore, according to the method and the device, the secret state domain model is trained in the TEE, so that the private data can be safely encrypted out of the domain and cannot be restored in a plaintext computing power environment, and the protection of the private data of the client is realized. Meanwhile, the large model is trained through the centralized outsourcing computing power of the server side, and it is guaranteed that centralized training is efficient.

The invention discloses a multi-source federal cross-domain and source-domain enhanced millimeter wave action recognition method and system, and the method comprises the steps: generating a micro-Doppler spectrogram through a millimeter wave radar, and extracting the motion features of human body actions through a signal processing module; in the federal multi-source domain adaptation module, dynamically evaluating and fusing knowledge of a plurality of source domains by adopting a voting-based pseudo-tag method and a weighted knowledge aggregation mechanism, and optimizing the generalization ability of a target model; and through a generalization gap optimization method, the performance of the source domain model is improved, and the robustness of the system in different environments is ensured. Through combination of a federated learning framework and a multi-source domain adaptation technology, unsupervised learning under the condition that a target domain has no annotated data is realized, only a single set of millimeter wave equipment is needed, a millimeter wave communication protocol is compatible, and the method has the characteristics of privacy protection, unsupervised learning, multi-source knowledge fusion and strong generalization ability. The method is suitable for application scenes of smart home, health monitoring, man-machine interaction and the like, and has wide practical application value and research prospect.

Systems and methods for domain-specific model-generated content item generation, evaluation, and selection can include generating a plurality of candidate model-generated content items that can then be evaluated based on one or more signals, which can then be leveraged for candidate model-generated content item selection. The plurality of candidate model-generated content items can be generated with a generative model that was tuned for domain-specific content item generation. The selected model-generated content item can be processed to generate an outline that may then be provided to a user for user interaction to generate an augmented outline. The augmented outline may then be processed to generate an updated model-generated content item.

The invention provides a medical record analysis method and device, electronic equipment and a readable storage medium, and the method comprises the steps: enabling a planning agent to divide each process step of a medical record analysis process into a plurality of minimum sub-tasks according to prompt words, and enabling the planning agent to distribute a corresponding task execution agent for each sub-task, the intelligent agent work assembly line executes the medical record analysis process, not only is the semantics of the complex medical context adapted, but also the efficiency and the accuracy of medical record analysis are improved, and based on the structure of the intelligent agent, the large language model module is a vertical domain model with professional domain task execution capability, and the large language model module is a large-language model with professional domain task execution capability. And meanwhile, the input and output of each node agent in the agent work assembly line are transparent, readable and interpretable, so that the stability of the output quality can be ensured.

The invention provides an entity relationship extraction method and device based on a transfer learning model, equipment and a medium, and belongs to the technical field of data processing. Original data are acquired, the original data are preprocessed and labeled, a training data set is generated, and then an entity relationship is constructed; utilizing a machine learning algorithm to construct an entity relationship processing task in the knowledge graph; constructing a transfer learning model, and carrying out transfer learning on knowledge in a source domain or a source task to a target domain or a target task based on the transfer learning model to initialize or assist training of a target domain model; a machine learning algorithm and a transfer learning model are fused, feature representation useful for a target domain can be extracted by utilizing rich annotation data and domain knowledge of a source domain, and the feature representation is applied to tasks of the target domain, so that useful features can be learned more quickly, and the efficiency of the target domain is improved. Therefore, the accuracy of tasks such as entity recognition and relation extraction is improved.

The invention discloses a cross-domain model parameter security aggregation processing system based on federated learning, and the system comprises a decentralized identity and reputation management module which is used for registering nodes and initializing reputation; the dynamic aggregation node election module is used for electing aggregation nodes from candidates based on reputation; the local security processing and proof generation module is used for generating encrypted model update verifiable proof at each node; the encryption transmission and on-chain verification module is used for verifying, proving and confirming the legal contribution; the trusted adaptive aggregation and execution proof module is used for safely aggregating the legal contribution in the trusted environment of the aggregation node, and generating a new model and execution proof; and the global state updating and reputation feedback module updates the reputation of each node according to the execution proof to complete a closed loop. According to the method, a decentralized model parameter aggregation framework is constructed by adopting a distributed consistency account book management system rule and combining with a dynamic aggregation node formed by a trusted execution environment, so that periodic rotation and hardware-level security isolation of an aggregation task executor are realized.

The invention relates to an uncertainty perception passive multi-target field adaptive image classification method, which is used for image recognition of autism spectrum disorder patients. According to the method, firstly, source domain model parameters are obtained and used for initializing a target model, then resting state functional magnetic resonance images of a plurality of imaging centers are preprocessed, and a plurality of target domains are constructed. On this basis, a current most representative target domain is selected through a minimum inter-domain difference strategy, an uncertainty modeling method based on evidence deep learning is adopted to train a target model, and class feature consistency is improved through domain contrast learning based on a class prototype in combination with a dynamically expanded auxiliary data set; and generating a pseudo tag to relieve the influence caused by tag noise. And finally, a trained target model is obtained through fine tuning optimization, and accurate classification of unknown images is realized. The method does not need to access source domain data, has the advantages of high robustness, high generalization ability and the like, and is suitable for actual cross-center medical image analysis scenes.

The invention discloses a self-adaptive visual admittance control method fusing fluid characteristics and multi-modal perception, which comprises the following steps: designing a self-adaptive Bingham-shear thickening fluid virtual damping coefficient through nonlinear mapping based on sigmoid, and combining a threshold triggering behavior of a Bingham fluid and a sudden stiffening characteristic under the impact of the shear thickening fluid; the flexibility is enhanced under the action of small force, and the anti-interference capability is improved under impact. Besides, a force auxiliary function based on force amplitude is introduced, an anisotropic compliance strategy is combined, rigidity and damping are dynamically adjusted by identifying the main force direction, and the mechanism can reduce sensitivity to noise of a micro sensor and ensure stability and accuracy in the task execution process. Meanwhile, an environment attraction domain model is established in a feature space, Lyapunov analysis shows that the system has consistent final boundaries, stable convergence is ensured, and secondary correction is supported.

The invention provides a transformer area distributed photovoltaic power prediction method and system, and the method comprises the steps: carrying out the clustering according to meteorological feature vectors in meteorological data through employing a clustering algorithm, dividing a source domain shared data set into similar data sets of different weather prediction scenes, training an Informer basic source domain model through employing a key representative sample in each prediction scene, and carrying out the prediction of the power of a transformer area. In the prediction process, the similarity between the meteorological characteristics of the day to be predicted and the meteorological characteristics of the typical weather scene is calculated, and the weather prediction scene and the corresponding basic source domain model are dynamically matched for each transformer area; and screening a key historical similar day of a day to be predicted of each transformer area as a micro training set, and dynamically updating decoder parameters of the adaptive Informer basic source domain model based on a model fine tuning thought of transfer learning to obtain a power prediction result of each transformer area. According to the method, the pertinence of a prediction result is improved, so that rapid modeling of distributed photovoltaic power prediction of massive transformer areas is realized.

The invention relates to the field of data annotation, and particularly discloses a data annotation method and system based on a large model, a terminal and a medium. Loading the fine-tuned domain model to carry out batch pre-labeling on the standardized data set to obtain a pre-labeling result of the standardized data set; calculating the prediction uncertainty of each sample in the standardized data set, and selecting a plurality of samples according to the prediction uncertainty to form a first to-be-audited data set; predicting the contribution degree of each sample in the standardized data set to the improvement of the domain model, and selecting a plurality of samples according to the improvement contribution degree to form a second to-be-audited data set; taking a union set of the first to-be-audited data set and the second to-be-audited data set to generate a to-be-audited target data set, and manually auditing the pre-labeling result of each sample in the to-be-audited target data set; and obtaining a labeling result of the standardized data set according to a manual auditing result. According to the invention, the data labeling efficiency and precision are improved.

The systems and methods disclosed herein relate to querying data using artificial intelligence models. A generalized model receives an output generation request and partitions it into segments mapped to specific domains, where each domain indicates associated databases and guidelines. The segments are routed to domain-specific models trained on domain-specific data, which generate query fragments by comparing performance metrics and system resource usage metrics. The query fragments are aggregated into an overall query that satisfies guidelines across domains. The systems and methods can include a feedback loop to adjust the domain-specific models using user interactions and performance metrics to dynamically adapt to a skill level or experience of the user.

The invention discloses an intelligent ship autonomous collision avoidance whole process path planning method, which comprises the following steps: establishing an error distribution model of a fusion target according to dynamic information and static information of a target ship, and analyzing to obtain a classification model of the target ship; according to the classification model of the target ship, combining a quaternary ship domain model to establish a ship domain model of dynamic and static obstacles; establishing local path planning based on an artificial potential field method, establishing a repulsive force field around an obstacle in the ship domain model, establishing a gravitational field at a target point, introducing a global path return mechanism, and establishing a local collision avoidance path planning algorithm considering the return mechanism; gravitational force self-adaptive adjustment is achieved based on the current ship path deviation degree and the current ship navigation deviation degree, a dynamically-optimized local collision avoidance path planning algorithm considering a self-adaptive homeward voyage mechanism is formed, and path planning of the ship is conducted. After collision avoidance of the ship, the ship is controlled to return to the global route reasonably and efficiently, and the navigation safety and efficiency of the ship are guaranteed.

The invention belongs to the technical field of knowledge Token generation and traceability, and discloses a knowledge Token generation and traceability method based on structured metadata, which is characterized in that a current knowledge scene is quickly matched and a core field is dynamically adjusted through few-sample learning in combination with various technical models and a preset industry field model library; a mixed semantic segmentation mode is adopted, knowledge fragments are completely segmented according to business logic, then a refined element extraction strategy is matched, three types of key elements including entities, relationships and clauses are accurately extracted, and errors are reduced through multi-dimensional verification; originally dispersed PDF, audio and video and other unstructured knowledge can be converted into standardized assets with anchoring information. Meanwhile, a multi-dimensional verification process is designed, comprehensive verification is carried out from a data source, access permission and logic consistency to a health state, it is ensured that a knowledge source can be checked, authenticity can be distinguished, and the information distortion risk during large model calling is reduced.

The invention provides a method and device for carrying out privacy protection on data in large model reasoning, and relates to the technical field of computers.The method comprises the steps that a large model is divided into a secret state domain model and a plaintext domain model, and the secret state domain model is placed in a trusted execution environment of a server side to be isolated; privacy data is encrypted by using a derivative key generated by the trusted execution environment measurement secret state domain model, and can be transmitted into the trusted execution environment; inferring the privacy data through a secret state domain model and a random confusion low-rank matrix connected in parallel with the secret state domain model, and transmitting inferred inference data to a plaintext domain model for continuous inference to obtain target inference data; and based on the client public key UserPK and the client private key UserSK, returning an inference result plaintext of the private data to the client. Therefore, large model reasoning is realized by means of centralized outsourcing computing power of the server, and privacy data is protected from being leaked in the reasoning process.

The invention relates to the technical field of model transfer learning, and discloses a model adaptive optimization method based on transfer learning. The method comprises the steps that source domain model structure parameters and target domain task initial data distribution are obtained, the feature mapping relation of all levels of a source domain model is extracted, and a cross-domain feature migration reference topological framework is generated; dividing a migratable feature layer and a to-be-reconstructed feature layer according to a target domain data distribution difference, and dynamically adjusting a migration priority in combination with a sample distribution density; freezing and unfreezing the transferable feature layer layer by layer based on the priority, synchronously constructing a local feature reconstructor, and optimizing domain offset through iterative feature alignment; collecting a feature reconstruction error and a migration feature retention degree in each iteration, and calculating a dynamic balance coefficient to adjust a freezing proportion and reconstruction intensity; and fusing the two types of features through a global model integrator, and generating mixed feature representation to drive end-to-end training of a target domain task.

The invention discloses a robot operation and maintenance long thinking chain corpus generation method based on a knowledge graph. The method comprises the steps of establishing an original corpus, generating a cleaning corpus, generating a deduplication corpus, performing intelligent blocking, generating question and answer pairs, performing entity recognition, performing entity mapping, determining a target reasoning path and generating a long thinking chain corpus. According to the method, a knowledge graph reasoning path is introduced as a generation constraint, so that model illusion is effectively inhibited, and explicit reasoning, logic completeness and full-link traceability of the reasoning process are realized; and meanwhile, in combination with puzzle-driven partitioning and domain model fine tuning, the semantic coherence and professional accuracy of the corpus are remarkably improved, and low-cost, large-scale and high-quality operation and maintenance corpus automatic production is realized.

The invention discloses a data statistical analysis method and device for intelligent questions and answers and a medium, and the method comprises the steps: obtaining a to-be-queried natural language, and determining an industry field where the natural language is located; in a domain model database, the industry domain is retrieved, and a target statement generation model is obtained; inputting the natural language into the target statement generation model to obtain an SQL query statement; and pulling query data from a statistical database by executing the SQL query statement, and performing statistical analysis on the query data. Data query is automatically and efficiently completed, and the accuracy of data query and statistical analysis is improved.

The invention discloses a real-time production data acquisition and intelligent scheduling method and system based on the Internet of Things, and relates to the technical field of artificial intelligence, an abnormal result is calculated based on a fusion weight and a predicted abnormal probability by constructing an abnormal detection model composed of sub-models, and when the abnormal result is greater than an abnormal judgment threshold, an abnormal state is judged; a threshold value self-adaption module is embedded in the anomaly detection model, and an anomaly judgment threshold value is output by taking minimization of the total misjudgment rate of anomaly detection as an optimization target; selecting a device and taking an anomaly detection model as a source domain model, migrating to a target device to obtain a fine-tuned target model, and predicting an anomaly probability value; identifying the key index, calculating the comprehensive health degree score of the key index based on the evaluation rule base, and judging the health state grade of the target equipment; and on the basis of the abnormal probability value and the health state level on the key index, the overall health score of the target equipment is judged, and intelligence and self-adaptability of equipment anomaly detection and health management are realized.

The invention provides an ethical examination method and device based on a large language model, and the method is characterized in that the method comprises the following steps: S1, carrying out the secondary pre-training, secondary instruction fine tuning and human feedback reinforcement learning of an existing large language model according to the existing training data, and obtaining a trained large language model as a domain model; s2, inputting the project document into an existing general model, and obtaining a preliminary review result corresponding to the project document in combination with a retrieval enhancement method and a review rule base; and S3, inputting the preliminary review result into the domain model, and obtaining a review result in combination with the review rule base. In a word, the method can assist the reviewer in rapid and efficient ethical review.

The invention discloses a textile workshop optimization algorithm recommendation method and system based on a large language model, and relates to the technical field of textile workshop scheduling, and the method comprises the steps: S1, constructing a textile knowledge vector database comprising a first-stage textile modeling knowledge base and a second-stage algorithm code knowledge base; s2, taking the spinning knowledge vector database as a fine tuning data set, and performing spinning field adaptive fine tuning on the large language model to obtain a spinning field model; and S3, based on the textile knowledge vector database, taking the obtained textile workshop scheduling demand as input, and using a retrieval enhancement generation method to generate a recommendation algorithm code for textile workshop scheduling. The textile field model is obtained by establishing the textile knowledge vector database and performing efficient fine tuning in the textile field on the large language model, the executable code is automatically generated through the retrieval enhancement generation method, the scheduling requirements and constraint conditions of different textile workshops can be adapted, the manual intervention cost is reduced, and the scheduling efficiency is improved. And the reusability and the scheduling scheme solving efficiency are improved.

The invention discloses a source domain irrelevant cross-domain cardiac beat recognition method and system for pseudo-label mining, and relates to the technical field of pseudo-label learning, and the method comprises the steps: obtaining electrocardiogram data with cardiac beat labels in a source domain, inputting a pre-established source domain model for pre-training, and obtaining a pre-trained source domain model; acquiring unlabeled electrocardiogram data of a target domain, and screening and classifying the data of the target domain based on a pre-trained source domain model and a preset category threshold to obtain data with high false label confidence and data with low false label confidence; using the source domain model to initialize target domain model parameters, revising a strategy based on a pseudo label of local and global semantic perception, updating the pseudo label of the low-confidence data, and combining the high-confidence data to obtain updated pseudo-labeled target domain data; target domain data subjected to data augmentation pseudo labeling are input into a target domain model, a target domain model optimization total loss function is calculated, and cross-domain cardiac beat intelligent recognition irrelevant to a source domain is achieved.