29 results about "Knowledge module" patented technology

The invention discloses an LLM central multi-agent tool arrangement and virtual-real closed-loop evolution scheduling system. The system comprises an intelligent sensing module, a knowledge module, an intelligent scheduling decision module and a closed-loop execution module. The intelligent sensing module collects real-time and multi-modal data of a physical environment and outputs structured sensing data, the knowledge module is used for constructing a knowledge graph and a historical task case library, and the intelligent scheduling decision-making module takes a large language model (LLM) as a core scheduling center, receives and analyzes task instructions, fuses the structured sensing data and graph and case information, and performs task scheduling on the knowledge graph and the historical task case library. Generating a decision scheme according with a constraint condition by using knowledge enhanced reasoning, and continuously optimizing in multiple iterations to generate a decision instruction; and the closed-loop execution module receives and drives the execution end to execute the decision instruction. According to the method, an intelligent closed-loop system of perception-decision-execution-evolution is constructed, and global optimization and high-reliability self-adaptive scheduling of multi-machine collaborative welding are realized.

The invention discloses a multi-agent debate-based content auditing optimization method, which comprises the following steps of: 1, acquiring and preprocessing multi-source content data: integrating to-be-audited content and associated context data, and constructing a comprehensive and reliable input data basis through standardization processing and quality verification; 2, intelligent agent role design based on domain knowledge: designing a specialized intelligent agent role in combination with audit scene characteristics, and injecting a domain knowledge module to enhance the understanding ability for complex semantics and rules; and step 3, a multi-agent debate and dynamic optimization framework: adopting a debate stream and arbitration stream parallel multi-agent debate optimization framework, mining deep semantics through multiple rounds of structured debate, and combining a dynamic arbitration mechanism to integrate multi-party viewpoints. According to the method, accurate identification, fair judgment and efficient processing of complex contents are realized through multi-agent structured debate and a full-process optimization mechanism.

The invention discloses an NL2SQL question and answer method and system based on an entity recognition small model and a large language model, and the system comprises an intention recognition module which is used for searching a related knowledge base according to a user question, inputting the large language model through constructing a cue word, judging the natural language intention of a user, and triggering a subsequent branch module; the entity recognition module is used for extracting key entities and related knowledge in a natural language of a user as components of cue words through a Bert model-based fine-tuning trained entity recognition small model in an NL2SQL model; the table selection module is used for vectorizing a natural language input by a user, retrieving keywords of a data table, matching similar questions and corresponding SQL examples in a sample library, and constructing cue words for a large language model to select a target data table; the large model cue word construction module is used for designing and optimizing a cue word template, and introducing a related knowledge module and a reference example module into a traditional role, task and constraint structure; the SQL statement execution and check error correction module is used for processing the generated SQL statement, if an error is reported, a new prompt word is constructed based on error information, a user problem and the SQL, check error correction is performed through a large language model, and execution is performed again for at most three times; according to the invention, more efficient and accurate NL2SQL service is provided for the user.

The application discloses a vehicle component detection method and system based on features and priori, and a medium. The method comprises the following steps: constructing a multi-target vehicle component dataset suitable for a vehicle loss assessment scene, completing data labeling and quality checking; configuring a hardware environment and a software framework required for training, and setting training hyperparameters; based on a convolutional neural network, integrating an enhanced feature pyramid network and an embedded component knowledge module to form a detection branch and a segmentation branch; training a detection network framework by using the dataset to obtain a trained detection model, and verifying the model detection and segmentation performance; and inputting a vehicle component image to be detected into the trained detection model to perform vehicle component detection. The application focuses on the guiding effect of knowledge on a deep learning model, enhances feature knowledge of a vehicle image, and introduces priori knowledge among vehicle components to improve the detection and segmentation accuracy of the model for multi-target vehicle components, and provides reliable technical support for vehicle intelligent loss assessment.

This application provides an intelligent dialogue Q&A method and system for operation and maintenance technical services based on a large model knowledge base. It relates to the field of large model operation and maintenance dialogue technology. The method involves receiving operation and maintenance consultation requests, capturing the dynamic evolution characteristics of the problem, and associating them with the real-time updated knowledge modules of the large model knowledge base to obtain corresponding relationships. Based on these relationships and phased descriptions, dynamic scenario adaptation processing is performed. An operation and maintenance Q&A generation module with knowledge iteration functionality is invoked, combining the adaptation results and incremental knowledge to generate multiple rounds of preliminary Q&A content. The preliminary Q&A content is dynamically optimized and adjusted based on feedback information to obtain dynamically optimized Q&A content. The content is fed back to the initiator, and relevant information is associated with the knowledge iteration module to drive incremental updates, completing the intelligent dialogue Q&A process and improving the accuracy and practicality of operation and maintenance Q&A.

The invention discloses a generalized quantitative perception training method and system based on knowledge feature distillation, and belongs to the technical field of intelligent computing. The method comprises the following steps: constructing a full-precision floating-point teacher model and a mixed bit width quantification student model based on a neural network architecture and a target task scene; carrying out pre-training on the full-precision floating point teacher model; taking the pre-trained full-precision floating-point teacher model as a knowledge transmission subject and the mixed bit width quantification student model as a knowledge receiving subject to carry out quantification perception training; the quantitative perception training is integrated with an attention feature fusion module and a multi-level context knowledge module; integrating a knowledge feature distillation strategy in the quantitative perception training process to realize knowledge extraction, migration and fusion of multilayer features; and a knowledge distillation loss function is calculated, the knowledge distillation loss function is taken as an optimization target, training optimization of the mixed bit width quantification student model is completed, and universal quantification perception training is realized.

The application discloses a cyclone intensity prediction method and device, and the method comprises the following steps: acquiring physical factor data, satellite cloud image data and cyclone intensity data in a historical period; performing pretreatment on the physical factor data, the satellite cloud image data and the cyclone intensity data to obtain a physical factor image, a satellite image image and historical cyclone intensity; inputting the physical factor image, the satellite image image and the historical cyclone intensity into a tropical cyclone intensity prediction model to obtain a 24-hour cyclone intensity prediction result; wherein the tropical cyclone intensity prediction model comprises a hierarchical fusion subnetwork, a deep convolutional neural network, a prior knowledge module and an uncertainty estimation module. The fusion of the uncertainty estimation and the long-term prior knowledge improves the prediction ability of the cyclone intensity change trend and the accuracy of the intensity prediction.

The invention provides an equipment end animal voice recognition model training method and system based on knowledge distillation, and the method comprises the steps: firstly carrying out the dynamic disassembly of knowledge of a pre-trained animal voice recognition teacher model, and obtaining a hierarchical knowledge module set which comprises basic features, correlation decisions and adaptive migration knowledge modules; acquiring real-time resource state information of the equipment side, and constructing a student model basic structure adapted to the equipment side; then establishing a layered associated distillation constraint rule to form a layered distillation constraint mechanism; inputting the animal sound training data set into a teacher model and student model basic structure, and executing dynamic distillation training in combination with a hierarchical distillation constraint mechanism to obtain knowledge transmission efficiency data and a model parameter change sequence; and finally, carrying out dynamic convergence adjustment on the basic structure parameters of the student model based on the data to obtain an equipment end animal voice recognition model. According to the invention, the training quality and efficiency of the equipment-side model are improved.

The invention provides an operation and maintenance technology service intelligent dialogue question answering method and system based on a large model knowledge base, and relates to the technical field of large model operation and maintenance dialogues, and the method comprises the steps: receiving an operation and maintenance consultation request, capturing the dynamic evolution characteristics of a question, and associating a real-time updating knowledge module of the large model knowledge base to obtain a corresponding relation; carrying out dynamic scene adaptation processing based on the corresponding relation and the stage expression information; an operation and maintenance question answering generation module with a knowledge iteration function is called, and multiple rounds of preliminary question answering content are generated in combination with the adaptation result and incremental knowledge; dynamically optimizing and adjusting the preliminary question answering content according to the feedback information to obtain dynamically optimized question answering content; and the content is fed back to the initiator, and related information is associated to the knowledge iteration module to drive incremental updating, so that the intelligent dialogue question answering process is completed, and the accuracy and practicability of operation and maintenance question answering are improved.

Methods of determining the species of birds in flight are provided along with corresponding systems. A method may include capturing a video stream of a bird in flight using at least one camera, generating a first species probability estimate by delivering images from the video stream to a neural network that has been trained to recognize species of birds from images, obtaining additional parameters from the video stream or from additional data, generating a second species probability estimate by delivering the additional parameters as input to a domain knowledge module with a domain knowledge statistical model, and generating a final species probability estimate by combining the first species probability estimate and the second species probability estimate. The additional parameters may include geometry features related to movement of the bird in flight, or parameters relating to the environment.

The invention discloses a simulation result processing system and method of a distribution network digitization system, and relates to the technical field of power distribution network digitization, and the system comprises a data module which is used for obtaining a fusion data packet based on simulation result data of a simulation model and power distribution network data; the feature module is used for inputting the fused data packet into a space-time diagram convolution feature extractor to obtain a high-order feature matrix; the causal module is used for inputting the high-order feature matrix into a causal discovery engine to obtain a dynamic causal graph; the mode module is used for obtaining an operation mode based on the high-order feature matrix and the dynamic cause and effect graph; the knowledge module is used for obtaining a mode evolution graph based on the operation mode and the dynamic causal graph and obtaining a knowledge base based on the mode evolution graph, and the problem that reusable systematic knowledge cannot be formed due to the fact that an existing simulation result processing system cannot automatically mine deep laws is solved.

This invention relates to a retrieval enhancement generation method based on multi-level knowledge fusion, belonging to the field of natural language processing technology. It aims to solve two core problems in existing retrieval enhancement generation technologies: fragmented knowledge levels and interference from noise information. The invention includes the following steps: Graph index construction: Segmenting documents into fragments and extracting entities and relations to generate a knowledge graph with key-value pairs, followed by deduplication and optimization; Multi-level knowledge retrieval: Acquiring local and global knowledge separately, and constructing cross-level semantic associations by connecting knowledge modules through an intermediate layer; Dual-dimensional filtering: Filtering core knowledge based on a weighted score of semantic relevance and knowledge importance, combined with a maximum boundary relevance strategy to remove redundant noise; Response generation: Inputting the filtered knowledge into a large-scale language model to generate accurate and coherent responses. This invention significantly improves the generation quality of knowledge-intensive tasks through cross-level knowledge bridging and precise noise filtering, possessing significant theoretical and practical application value.

The invention is suitable for the technical field of collection training, and provides a self-adaptive collection learning path planning method and system, and the method comprises the steps: constructing a three-dimensional procurement capability model containing compliance operation, a negotiation strategy and risk management and control, decomposing the procurement capability model into capability units, and setting a quantitative index; collecting operation, decision and assessment data in the simulation scene in real time, and generating a dynamic capability assessment matrix; constructing a three-dimensional space model based on the matrix, calculating the association strength and the lifting priority of the power unit, and outputting a personalized learning path; recombining the knowledge module according to the evaluation value change of the capability unit; dynamically adjusting model parameters in combination with policies and industries and re-planning a path; a cross-dimension learning path is generated when the plurality of capability units reach the standard. According to the method, dynamic evaluation, personalized path adjustment and cross-dimension capability integration can be realized, the problems of one-sided evaluation, fixed path and insufficient adaptability in the prior art are solved, and the accuracy and efficiency of auditory collection training are improved.

The invention discloses an electric energy quality disturbance identification method based on a constraint network and priori knowledge. According to the method, the constraint network and the prior knowledge are introduced, so that the recognition difficulty of the main network on multiple disturbance types is reduced, and meanwhile, the algorithm recognition accuracy is improved. The constraint network and priori knowledge model comprises a one-dimensional fusion convolutional network module used for feature extraction of an original disturbance signal; the non-parameter attention mechanism optimized long and short term memory network module is used for capturing time sequence characteristics of power quality disturbance; the constraint network module is used for identifying the disturbance quantity contained in the power quality disturbance signal; and the priori knowledge module is used for correcting the classification result of the conflict labels. According to the method, the main network is assisted by the constraint network, the complexity of the main network is reduced, the recognition accuracy is improved, and the method has the advantages of light weight and high precision and has good practical application potential.

The invention belongs to the technical field of aquaculture, and particularly relates to an intelligent river crab culture monitoring system and method. Comprising a breeding system sensing module, a breeding system feature analysis module, a breeding technology knowledge module and a decision efficiency evaluation and feedback module. The method has the advantages that production data, including water quality, weather, water color, aquatic plants and crab individual characteristics, of the whole breeding process are completely recorded, and feeding, drug use records, water changing, shelling, aeration and other records in the breeding process are recorded. And a data basis is provided for follow-up green aquatic product traceability, rating and culture technical guidance. According to the method, breeding behavior effect quantification and next decision optimization are carried out, effects generated by specific breeding behaviors are evaluated from apparent characteristics of water quality, water color, aquatic plants and crabs, and the effects are added into breeding technology knowledge, so that richer experience knowledge is provided for next decision.

The application provides a supervision multi-agent modular collaborative architecture system and method fusing a large model, and belongs to the technical field of artificial intelligence supervision.The method comprises the following steps: constructing a scenario-based agent resource capsule, integrating a structured supervision knowledge module, a supervision rule engine module and a supervision tool service module; deploying an audit supervision scenario-based intelligent engine based on the scenario-based agent resource capsule; and establishing a private security collaborative framework, wherein the private security collaborative framework comprises a security isolation mechanism, a permission hierarchical control module and a security log audit mechanism.The supervision multi-agent modular collaborative architecture method fusing the large model significantly improves the efficiency and flexibility of audit supervision; through intelligent scheduling of supervision resources, the time consumption and error rate of manual context splicing are greatly reduced, and the accurate output of the supervision logic is ensured.

The invention relates to the technical field of file process management, in particular to a file process management method and system based on dynamic distribution and consensus tracking, and the method comprises the steps: receiving an original file project voucher and process starting metadata, carrying out the project scene context research and judgment and context correlation analysis, and generating a project scene feature list; generating a standardized project flow path through project flow path research and judgment and a multi-target project optimization strategy; initiating a project consensus verification item at each project consensus verification node; arranging and forming a project consensus verification summary, and performing sequential archiving by adopting a project evidence storage management specification to form a project consensus tracing file; and performing project process efficiency research and judgment and project knowledge value extraction based on a project consensus tracing file, generating a project process redisk report, organizing a project knowledge module, and feeding back and updating. According to the method, dynamic intelligent planning of a file flow path, credible solidification of a consensus process and closed-loop multiplexing of flow knowledge are realized.

This invention relates to an immersive virtual simulation teaching system for chemistry courses, belonging to the technical field of educational systems. It includes: a chemistry knowledge module for storing professional knowledge content of the chemistry course; an element library module for storing educational resources related to chemistry knowledge; an immersive interaction module for constructing three-dimensional virtual chemistry experiment scenarios; and an evaluation and feedback module for collecting user operational behavior data, decision-making path data, and reflection text data during the virtual simulation experiment process, generating a user competence evaluation report. This immersive virtual simulation teaching system for chemistry courses, by constructing a virtual simulation teaching system including a chemistry knowledge module, an element library module, an immersive interaction module, and an evaluation and feedback module, achieves for the first time a deep immersive integration of chemistry professional education and general education, transforming education from passive acceptance to active experience and value internalization.

An internet of things safety auxiliary system based on large language model assistance comprises an equipment identification module, a safety knowledge module and a safety management suggestion module, equipment identification and equipment safety suggestion providing are integrated, safety control over internet of things equipment is enhanced comprehensively and automatically, the text analysis and generation capacity of LLM is utilized, and the safety of the internet of things is improved. The RAG technology is used for providing support for a user, and the Internet of Things flow triggers response of the safety auxiliary system after being identified by equipment.

An automatic personalized response generation system and process receives user email containing questions and knowledge base articles through an email receiver integrated to an email engine and stores the user emails in a knowledge database. A gatekeeper module classifies incoming emails using structured criteria. A knowledge module gathers and organizes relevant information from the knowledge database and the gatekeeper module. An orchestrator module then routes the classified emails and relevant information to either a responder module or a human agent. The responder module crafts personalized responses, which are then refined by a modify module according to specific quality criteria. Throughout the personalized response generation process, the gatekeeper module, the knowledge module, the orchestrator module, the responder module, and the modify module utilize specialized prompts to guide an Artificial Intelligence (AI) engine via a prompt generator in performing its designated tasks. The email engine transmits the modified personalized responses back to user.

The invention provides a medical image segmentation system based on knowledge guidance and prompt generation, which relates to the technical field of medical image segmentation and comprises a data acquisition module, an expert knowledge base, a knowledge scheduler, a parallel knowledge calculation unit, a prompt fusion device and a universal segmentation model with a prompt encoder. The expert knowledge base comprises a plurality of expert knowledge modules and is used for generating rich information prompts; the knowledge scheduler is used for selecting expert knowledge modules from the expert knowledge base; the parallel computing unit simultaneously executes an expert knowledge selection module; the prompt fusion device determines an optimal aggregation prompt set based on rich information prompts output by each expert knowledge module; and the universal segmentation model with the prompt encoder segments the medical image to be segmented into a corresponding segmented medical image based on the optimal aggregation prompt set. According to the method, the calculation of domain knowledge is separated from the segmentation capability of a general model, so that the problem of the general model in an automatic clinical process is solved.

The invention provides an atrial fibrillation intelligent auxiliary diagnosis and treatment system and method fusing a large language model and multi-task learning, a terminal and a medium. Through the arrangement of the intelligent core and knowledge module, the intelligent perception and fusion module and the intelligent analysis and decision engine, the semantic comprehension ability of the large language model and the logical reasoning ability of the knowledge graph are fused, medical contexts can be understood, clinical laws can be learned, and dynamic intelligent prediction and decision assistance are provided.

The invention discloses a multi-modal knowledge extraction and association method and system based on recorded and broadcast class content, and relates to the technical field of artificial intelligence, and the method comprises the following steps: obtaining an audio stream of a recorded and broadcast class, carrying out the voice recognition processing, obtaining a multi-modal text stream, constructing a multi-layer knowledge structure based on the multi-modal text stream, and carrying out the multi-layer knowledge extraction and association. The method comprises the following steps: establishing a multi-dimensional content file of microscopic knowledge atoms, constructing a global intelligent search engine based on a multi-level knowledge structure and the multi-dimensional content file, responding to a query instruction of a learner through the global intelligent search engine, and carrying out associated content recommendation. According to the method, a multi-layer knowledge system combining macroscopic chapters, mesoscopic knowledge modules and microscopic knowledge atoms is constructed, multi-modal information is deeply fused, a multi-dimensional content file is established for the microscopic knowledge atoms, and the knowledge graph is constructed based on the multi-dimensional content file, so that the dependency relationship and semantic association between knowledge points are effectively revealed, and the knowledge graph is established. And finally, the learning efficiency and the personalized learning experience are remarkably improved.

A dialogue system for cognitively impaired adults based on CST principles is provided. The dialogue system includes an encoder module, a decoder module, and an external knowledge module. The encoder module is configured to classify utterance and execute a progressive mask context training part and a text classification part. The progressive mask context training part performs a sentiment analysis function and a keyword extraction. The text classification part performs text classification tasks comprising emotion, strategy, and CST classifications. The decoder module is in electrical communication with the encoder module for response generation according to predicted values derived from attention weights in response to outcomes of the encoder module from the utterance. The external knowledge module, the encoder module, and the decoder module interact with each other to form a dialogue framework embedded with CST principles through multi-source knowledge fusion and to guide the dialogue framework to generate responses.

The application provides an AR-enabled traditional Chinese medicine culture table game card design method, relating to the field of augmented reality technology. The AR-enabled traditional Chinese medicine culture table game card design method comprises the following steps: S1: based on the theory of yin and yang, five elements, four qi and five tastes, and meridian and zang-fu organ, abstract traditional Chinese medicine knowledge is disassembled into independent knowledge point units through a modular extraction method to generate a standardized knowledge module library; S2: based on the standardized knowledge module library, a decision tree model is used to match traditional Chinese medicine theory and game logic, the five-element generation-restriction relationship and the syndrome differentiation and treatment process are converted into win-lose determination rules, and a theory-rule mapping table is generated. By disassembling abstract traditional Chinese medicine theory into independent knowledge point units and generating a standardized knowledge module library, the knowledge processing efficiency and subsequent technology adaptability are significantly improved. By using a decision tree model to convert the five-element generation-restriction relationship and the syndrome differentiation and treatment process into game win-lose rules, the lossless translation of traditional Chinese medicine logic and entertainment mechanisms is realized.

The invention provides a knowledge migration method and system supporting a multi-modal task, and relates to the field of artificial intelligence task migration. Obtaining source task data and a target task demand input by a user; extracting a source feature knowledge set of the source task data based on a target task demand, and performing cross-modal alignment to obtain a standardized knowledge module; analyzing the target task demand to obtain a target task feature; obtaining an initial knowledge migration strategy based on the current strategy library and the target task features, inputting the initial knowledge migration strategy into a pre-constructed strategy optimization agent, and outputting an optimal knowledge migration strategy; performing knowledge migration-oriented local fine adjustment on the basic model based on the optimal knowledge migration strategy and the standardized knowledge module to obtain an initial target task model; and detecting and updating the initial target task model through a pre-constructed evaluation index system to obtain a target task model. The reusable knowledge is migrated through the optimal knowledge migration strategy, and migration efficiency and performance are guaranteed.

The invention provides an industry knowledge data processing method and device and related equipment, and belongs to the technical field of artificial intelligence. The industry knowledge data processing method in the embodiment of the invention comprises the steps that according to a standardized knowledge structure of a plurality of knowledge modules containing a target industry, the target industry is obtained; creating a plurality of first intelligent agents of the plurality of knowledge modules and first cue words of the plurality of knowledge modules; obtaining an original document of the target industry, respectively inputting the original document and the first prompt words of the plurality of knowledge modules into the plurality of first intelligent agents to generate knowledge data, and obtaining structured knowledge data of the plurality of knowledge modules; summarizing the structured knowledge data of the plurality of knowledge modules according to the standardized knowledge structure to obtain stock knowledge data; and performing vectorization processing on the stock knowledge data to obtain vectorized knowledge data, and storing the vectorized knowledge data in a vector database.

The invention discloses a model fusion method based on dynamic combination modularization expert knowledge. The method comprises the following steps: 1) knowledge modularization; knowledge is divided into a shared part and an exclusive part, and common knowledge useful for all tasks is separated; and 2) dynamically merging the shared knowledge and the specific task knowledge based on input to obtain a fused model. The invention provides a method for merging models of a plurality of specific tasks into a single multi-task model without additional training.

The invention relates to the field of medical image technology and artificial intelligence cross technology, and discloses an ultrasonic intelligent collaborative quality control system. Comprising a data security module, a user login module, an artificial intelligence data large model, an automatic quality control selection module, a real-time online collaborative consultation module, a standard section interface module, an image acquisition and transmission module, a standard section classification module, a semantic segmentation module, a measurement index module and an anatomical knowledge module. The clinical guide module and the industry expert consensus reference management module are used for analyzing ultrasonic images by using an artificial intelligence large model, and optimizing the process through the automatic quality control module to ensure the image quality; the semantic segmentation module calculates the area ratio Qa of the lesion region and the three-dimensional volume ratio Qb of the lesion region, and provides an accurate diagnosis basis; the real-time online consultation supports multi-expert remote cooperation, the diagnosis efficiency is improved, the lesion segmentation precision evaluation unit monitors the system in real time to predict the lesion segmentation precision Px, the quality control process is automatically triggered when the lesion segmentation precision Px is lower than a threshold value, and it is guaranteed that the result is reliable.