82 results about "Knowledge utilization" patented technology

The invention discloses a traffic large model construction and decision-making method and device based on multi-modal two-way map reasoning, and the method comprises the steps: constructing a multi-modal data set of a text, an image and a track, generating fusion features through spatial-temporal clustering and cross-modal Transform coding, carrying out the two-way map reasoning in combination with a traffic knowledge map, and carrying out the decision-making of the traffic large model. The method comprises the following steps: generating an embedded representation through a forward graph neural network, reversely mapping a decision scheme generated by a language model to a graph to verify consistency, outputting knowledge to enhance embedding, fusing multi-modal features and knowledge embedding by adopting an LoRA multi-task joint fine tuning technology, adapting to traffic field tasks, deploying a real-time inference engine, and carrying out real-time inference on the traffic field. And processing the dynamic data flow through an aging perception attention mechanism, and outputting traffic event identification, path planning and scene question and answer results in parallel. Compared with the prior art, the method has the advantages that the problems of insufficient multi-source heterogeneous data fusion, low knowledge utilization efficiency and poor real-time decision consistency can be solved, and the semantic understanding and decision accuracy of the traffic large model is effectively improved.

The invention discloses a federal large model knowledge collaborative training method supporting multi-modal heterogeneous clients, which comprises the following steps: each client receives a model initialization parameter issued by a central server, and applies adaptive differential privacy noise to independently train a heterogeneous lightweight model based on local private data; updating the model to which the noise is applied and uploading a modal identifier of the model to a central server side; after model updating and modal identification of each client are received, based on a modal perception weighted consensus fusion mechanism, knowledge of each client is fused to update a global large model; and the central server side issues the updated presentation layer parameters of the global large model to the client side for initialization of the next round of local training. According to the method, on the premise that a public data set or specific task setting is not needed, comprehensive compatibility of data isomerism, client dynamic participation, model diversity and privacy protection requirements is achieved, and the adaptability, stability and knowledge utilization efficiency of large model federation training are remarkably improved.

The invention provides a large language model knowledge preference alignment method and system based on self-supervised learning. The method comprises the following steps: defining knowledge availability comprising three dimensions; generating and constructing preference data based on retrieval enhancement, obtaining constructed data according to the generated questions, answers and context paragraphs, evaluating the data through a defined knowledge utilization degree, and screening out samples for training; and training samples through a preference optimization method comprising a self-supervision module and knowledge perception to perform preference alignment, wherein the self-supervision module integrates knowledge utilization degree as a supervision signal into model training. Based on the technical scheme of the invention, the data acquisition cost is effectively reduced, the problems of scarcity of field data and difficulty in labeling are solved, and the data quality is improved; and the understanding and application capability of the model on professional knowledge is improved, and effective alignment of the model and domain expert knowledge application is realized.

The invention relates to a knowledge graph enhanced multi-modal file retrieval method, and belongs to the field of artificial intelligence and multi-modal information retrieval. The invention aims to solve the problems of multi-modal information semantic segmentation, weak semantic reasoning ability and low semantic matching precision in the existing multi-modal archive resource retrieval process. Through four stages of archive multi-modal data preprocessing and feature extraction, knowledge graph construction and enhancement, semantic retrieval request analysis and intention modeling, and multi-modal semantic matching and sorting, a semantic relationship is enhanced by utilizing a knowledge graph, so that the semantic relevancy and context consistency of a retrieval result are remarkably improved; a user is allowed to input and inquire in various forms such as texts, images and voices, and semantic extension retrieval is supported. According to the method, more efficient and accurate archive resource retrieval can be realized, and the user retrieval experience and archive knowledge utilization are improved.

The invention discloses an assembly process knowledge graph construction and updating method based on graph attention network mapping, and the method mainly comprises the following steps: standardization system design construction, data collection, data classification, multi-modal data preprocessing, multi-source information extraction and graph node feature initialization. Semantic modeling and vector space mapping driven by a graph attention network, knowledge storage, construction of an assembly process knowledge graph by using stored knowledge data, and semantic complementation and dynamic updating of a process information model. Through the unique self-attention mechanism of the GAT, the weight can be dynamically distributed for the relationship among different information nodes, so that knowledge reasoning and expression are more accurately carried out, and the problems of knowledge fragmentation, management rigidity, semantic understanding deficiency, low knowledge utilization efficiency and the like in the prior art are systematically solved.

The invention discloses a power grid equipment health management and operation and maintenance platform based on knowledge federation and a language model, and relates to the technical field of power grid equipment health management and intelligent operation and maintenance. Comprising a multi-source data fusion module, a knowledge federal management system, a large language model, an intelligent analysis and prediction module, an operation and maintenance decision support module and a visual display module, and the large language model can be a DeepSeek-R1 large model. According to the method, the knowledge federation technology and the breakthrough AI capacity of the DeepSeek-R1 large model are creatively fused, remarkable technical innovation and function breakthrough are achieved in the aspects of data fusion, knowledge utilization, man-machine interaction, intelligent decision making and the like, and the limitation of an existing power grid operation and maintenance technology can be effectively overcome; the intelligent and active preventive level of operation and maintenance of power grid equipment is comprehensively improved, and key technical support is provided for constructing a safer, more reliable and more efficient intelligent power grid.

The invention discloses a role playing model fusion system and method based on retrieval and reflection, and belongs to the technical field of natural language processing. The implementation method comprises the following steps: 1, performing data cleaning on original role historical dialogue data, and performing semantic screening by using semantic similarity to form historical dialogue context data; 2, event knowledge associated with the roles is constructed through association degree matching; 3, constructing a role reflection strategy mode by using historical dialogue data, and generating role reflection data; 4, when a user inputs a question, replying by adopting a multi-source fusion mode based on role thinking; compared with the prior art, in a multi-scene task, a dynamic retrieval role knowledge base is utilized to screen task related background knowledge, and a thinking mode modeling framework based on a role reflection mechanism and a context fusion strategy are constructed to generate high-quality answers conforming to the context; and thus, the fusion ability of the role playing model is improved, and the consistency of role words and language styles is realized.

The invention relates to a converter steelmaking knowledge graph construction method and device and computer equipment, and the method comprises the steps: constructing an ontology hierarchy according to collected multi-source heterogeneous data, then carrying out corresponding knowledge extraction on different types of data based on an ontology hierarchy structure, employing a similarity calculation hierarchical clustering method based on TF-IDF and SBERT, and carrying out knowledge extraction on different types of data; and converter steelmaking knowledge is processed and fused, so that the data quality of the knowledge graph is ensured. Multi-source knowledge dispersed in a technical document, an operation manual, an enterprise informatization database and the like is systematically integrated to form a structured knowledge base, which has positive influence on improving information retrieval and knowledge utilization and acquisition processes of converter steelmaking and provides solid knowledge data support for subsequent implementation of converter intelligent steelmaking.

The invention discloses a pollutant toxicity assessment method and device based on RAG enhancement and knowledge graph reasoning, equipment and a medium, and relates to the technical field of pollutant assessment, and the method comprises the steps: constructing a standardized toxicological knowledge graph through multi-source data collection and preprocessing; by combining RAG enhanced retrieval and knowledge graph reasoning, a toxicity reasoning path and an evaluation index are rapidly generated, meanwhile, interpretable content is provided, multi-source heterogeneous data fusion and self-adaption to spectral characteristic differences of different sites are achieved, the problems that traditional evaluation is low in efficiency, high in cost, insufficient in knowledge utilization and the like are effectively solved, rapid and accurate evaluation is achieved, and the evaluation efficiency is improved. And the credibility and the interpretability of the evaluation result are enhanced.

The invention relates to a knowledge graph construction and retrieval optimization method based on water affair data, and the method comprises the steps: taking the unified knowledge expression of a water affair object and an event thereof as a target, determining a water affair business field and business range, and constructing a water affair knowledge graph conceptual model covering a static knowledge ontology and a dynamic event ontology; performing knowledge extraction on the obtained multi-source heterogeneous water affair data, converting the data into a unified triple form, and constructing a data layer of a water affair knowledge graph; and on the basis of the data layer, node, path and region query is performed on the knowledge graph by adopting a graph algorithm, and semantic enhancement retrieval and reasoning are performed in combination with a large language model. By constructing the water affair knowledge graph and introducing knowledge extraction and retrieval enhancement capability of a large language model, dynamic modeling and semantic reasoning of water affair objects, events and relationships thereof are realized, high-risk area identification, emergency path planning and emergency early warning are supported, and the knowledge utilization rate and intelligent decision level in the water affair field are improved.

The invention relates to the technical field of big language model reasoning, and provides a robust retrieval enhancement method and device oriented to a big language model. According to the method, semantic analysis is carried out on user query in a real scene, a key entity is positioned, then a core entity is taken as an initial retrieval point, a credible iterative retrieval strategy is executed, associated information in a knowledge graph is obtained step by step, factual verification is carried out, the limitation of a traditional RAG framework on the key entity in the real scene is broken through, and the reliability of the key entity in the real scene is improved. Robust reasoning and knowledge utilization of the LLM under the conditions that reliable predefined associated entities are lacked and query semantics are complex are achieved, and robustness and fact consistency of the model in a complex knowledge graph question and answer task are improved.

The invention discloses a retrieval enhancement generation method and system based on long-term memory and multi-source knowledge iterative fusion. The method comprises the following steps: firstly, after user inquiry is received, triggering long-term memory database retrieval, external knowledge base retrieval and large language model internal knowledge generation in parallel to construct a multi-source candidate knowledge base, and improving result correlation by adopting a self-adaptive retrieval method based on a dynamic top-k value; secondly, inputting user query and multi-source knowledge into the large language model for iterative fusion, and generating a final answer through conflict detection, fusion processing and confidence evaluation; and finally updating the long-term memory database. The method can improve the accuracy and stability of the generated answers in a multi-round interaction and complex task scene, reduces the risk of answer one-sidedness, inference incompleteness or fact deviation, enhances the multi-source knowledge utilization ability and long-term learning ability of the system, and improves the user experience. Therefore, the comprehensive performance and reliability of the system in applications such as intelligent question answering, information retrieval assistance and text generation are obviously expanded.

The invention provides an electric power complex question-answering method and device based on large-model multi-agent, and relates to the technical field of electric power natural language processing. The method comprises the steps that a domain knowledge base is constructed, and a large-model multi-agent collaborative reasoning framework is constructed and comprises a service processing agent, a domain knowledge agent and an action planning agent; acquiring a natural language problem of a target electric power complex task input by a user by using a business processing agent; interaction between the business processing agent and the action planning agent is utilized to obtain action planning knowledge of the natural language problem; the business processing agent interacts with the domain knowledge agent, and domain knowledge of the target electric power complex task is obtained from the domain knowledge base; and generating an analysis result of the target electric power complex task in combination with the action planning knowledge and the domain knowledge in the business processing agent. According to the invention, the accuracy and reliability of power complex problem results can be improved.

The invention relates to the technical field of medical informatization, in particular to a clinical aid decision-making optimization method, device and equipment based on guide knowledge, and the method comprises the steps: obtaining electronic medical record data from a hospital information system and a laboratory information management system communication system, data standardization is achieved through the steps of duplicate removal, missing value processing, abnormal value processing, format standardization and the like; according to authoritative knowledge such as clinical pathways, diagnostic guidelines and disease consensus provided by doctors, key feature information is extracted by utilizing a natural language processing technology, structured data is extracted by adopting a regular expression, and unstructured texts are marked by pre-training a medical word vector dictionary, performing word segmentation by a classifier and performing part-of-speech tagging. The weight is optimized based on the loss function; an initial case knowledge base is constructed, and high-quality cases are screened out according to evaluation indexes such as knowledge richness and curative effect; dividing the high-quality cases into well-known physician cases and rare disease cases through syntactic analysis according to the classification indexes; case matching is carried out by adopting selective weighted heterogeneous value distance measurement, discrete attributes adopt value difference measurement, continuous attributes adopt standardized difference square, and attribute weights are optimized by means of a genetic algorithm, so that auxiliary clinical decision and knowledge services are realized. The invention can provide a scheme which can efficiently construct a high-quality case library, automatically extract key feature information, fuse guide knowledge for classification and screening, and introduce more accurate measurement and weight optimization in matching.

The invention discloses a vertical domain intelligence analysis method based on a large language model. The method comprises the following steps: receiving a natural language analysis demand of a user; based on pre-constructed vertical domain specific task planning knowledge, intelligently decomposing the demand into a subtask list corresponding to a preset dimension by using a large language model, and semantically matching a proper information acquisition tool for each subtask to generate a structured task plan; the system schedules and executes tool sets of network search, webpage capture, internal query and the like according to a plan, and performs cross verification and cleaning on collected fragmented information; and inputting the verified data into a large language model for integration analysis, and optionally carrying out multi-round iteration self-evaluation optimization to automatically generate a high-quality vertical domain intelligence analysis report. According to the method, full-process and high-intelligence automatic processing of information analysis in the vertical field from demand understanding to report generation is realized, and the method has excellent planning capability, process interpretability, result reliability and excellent modular expansibility.

The invention provides an autonomous vulnerability analysis system and method based on knowledge graph guidance and hierarchical verification, and relates to the technical field of automatic software vulnerability analysis, and the method comprises the steps: constructing a structured vulnerability knowledge graph by using unstructured historical vulnerability cases; obtaining a to-be-analyzed project, inputting the to-be-analyzed project into the large language model, taking the structured vulnerability knowledge graph as priori knowledge, and performing information retrieval by using a macroscopic mode and statistical evidence in the structured vulnerability knowledge graph; guiding a large language model to identify a pollution source and a risk pool based on a retrieval condition, and driving a static analysis engine to generate a preliminary suspected path; and screening the suspected paths through a two-stage grading review mechanism, finally outputting to obtain a high-credibility vulnerability, and generating a final high-credibility vulnerability report. The method is based on statistical law inference, so that the false alarm rate can be greatly reduced, and the discovery precision of real vulnerabilities is remarkably improved.

The application discloses a kind of urban land remote sensing image classification methods based on evolution multitask optimization, comprising: 1: obtaining urban land cover remote sensing image sample and pre-processing, to build sample dataset;2: the sample dataset is decomposed into multiple training subsets, and the corresponding sub-population is constructed for each training subset;3: the state of sub-population is evaluated to adjust the probability of executing knowledge transfer operation online;4: different source populations are selected according to the state of sub-population to provide knowledge for it;5: more optimal classifier is obtained by using the model representing global information;6: select a part of poor performance classifier to carry out unbalanced adjustment;7: finally, the best individual is selected from the sub-population to form a multi-classifier, to identify different types of urban land use and cover situation.The application can quickly classify different urban land types by constructing and training the optimal multi-classifier, so as to improve the accuracy and precision of land type classification result.

The application discloses a pollution toxicity evaluation method and device based on RAG enhancement and knowledge graph reasoning, equipment and medium, relates to the technical field of pollution evaluation, and comprises the following steps: constructing a standardized toxicology knowledge graph through multi-source data acquisition and preprocessing. Combined with RAG enhanced retrieval and knowledge graph reasoning, the toxicity reasoning path and evaluation index are quickly generated, the explainability content is provided, the multi-source heterogeneous data fusion and the adaptive difference of different site spectrum characteristics are realized, the problems of low efficiency, high cost, insufficient knowledge utilization and the like of the traditional evaluation are effectively solved, the rapid and accurate evaluation is realized, and the credibility and explainability of the evaluation result are enhanced.

The application discloses a kind of based on measurement and knowledge water surface unmanned system operating state early warning method, belong to unmanned system safety monitoring technical field.To solve the problem of insufficient knowledge utilization, poor dynamic adaptability of existing early warning method, the method of the application includes: collecting position, attitude, ship body state and environment multi-source time series data, and obtaining text form expert knowledge;Adopt sliding window and standardization preprocessing data;Expert knowledge semantic vector is extracted using BERT model;Through Crossformer model coding time series features, knowledge vector is spliced and fused, and system state features are output;Risk membership degree is calculated based on fuzzy set theory, and early warning is triggered according to the maximum membership principle.The application realizes the deep integration of measurement data and expert knowledge, improves the early warning accuracy, robustness and generalization ability.

The invention discloses a knowledge management method and system based on document similarity detection and a multi-modal knowledge base. The method comprises the steps of obtaining a target document; when a target document is put in storage, a similarity detection method based on a resource self-adaption mechanism is adopted, and existing documents are screened out from the multi-modal knowledge base to serve as a candidate document set; a similarity detection method based on a precision and efficiency optimal mechanism is adopted, and the candidate document set is used for conducting duplicate removal on the target document or storing the target document which is judged to be a new document; obtaining a question asked by the user; based on the multi-mode knowledge base subjected to duplicate removal or updating, multi-path mixed retrieval is conducted on the problem, a multi-path mixed retrieval result is subjected to two-stage sorting to obtain a rearrangement result, reference data are selected from the rearrangement result, the problem, cue words, context and the reference data are assembled into a large model input instruction according to a preset cue word template, and the large model input instruction is input into the multi-mode knowledge base. An answer generated by the large model is obtained and returned to the user; and comprehensive improvement of the knowledge utilization rate, the retrieval precision and the efficiency is realized.

The invention discloses a ship equipment fault detection method and device, a medium and a product, and relates to the field of electric digital data processing, and the method comprises the steps: introducing a bidirectional gating circulation unit and a variational auto-encoder into a teacher-student network model, and obtaining an initial detection model; respectively inputting a fault-free signal in the first sample data set and a corresponding signal in the second sample data set into a student network and a teacher network, training an initial detection model to obtain a ship equipment detection model, and inputting a to-be-detected ship equipment signal to obtain an actual detection result; in the training process, a BM method and an extreme value theory are adopted to obtain GEV distribution so as to determine a fault detection threshold value, and then ship equipment fault detection is achieved in combination with an actual detection result. According to the method, the accuracy, the robustness and the practicability of detection are remarkably improved, the technical bottlenecks in the aspects of data dependence, priori knowledge utilization, detection precision and the like in the prior art are broken through, and the detection requirements of complex and changeable statistical characteristics in a real environment are met.

This invention provides a sepsis consultation method and system based on multi-turn dialogue, belonging to the field of intelligent consultation equipment technology. This invention achieves effective integration of standardized medical knowledge and personalized clinical experience through the construction of a two-dimensional knowledge graph and cross-graph interactive retrieval, improving the accuracy and reliability of sepsis diagnosis and treatment recommendations. Through structured sepsis diagnosis and treatment chain constraints and a confidence-driven dynamic strategy, it alleviates the illusion problem of large language models in complex clinical reasoning, ensuring that the generated content conforms to clinical diagnosis and treatment logic norms. Through semantic anchor guidance, hierarchical activation, and knowledge feedback mechanisms, it achieves multi-granular semantic association from macro-clinical scenarios to micro-medical concepts, balancing retrieval accuracy and knowledge coverage. The confidence evaluation system comprehensively considers knowledge sufficiency, redundancy, and consistency, adaptively balancing knowledge utilization and information exploration, improving the reasoning efficiency and diagnosis and treatment quality of multi-turn dialogue.

The disclosure provides a sequence retrieval and classification method based on domain rule knowledge and data. It includes a key text extraction module, a specification and chapter structure tree module, and a loss correction module. The key text extraction module inputs the external text information after artificial review and the text knowledge in the domain rule knowledge specification and chapter, and obtains the key text sequence of each document. The specification and chapter structure tree module inputs the text knowledge in the domain rule knowledge specification and chapter, and uses the inherent hierarchical structure of the specification and chapter to which the document belongs to guide the retrieval process. The loss correction module calculates the loss by comparing the difference between the predicted document identifier and the true labeled document identifier. This algorithm uniformly minimizes the difference between the predicted and labeled classification / retrieval document identifiers, improving the accuracy and consistency of the two tasks.

The application discloses a colorectal cancer HER2 state classification method based on a double-branch cooperative optimization network and belongs to the technical field of pathological image processing. The application adopts a double-path cooperative optimization framework, combines knowledge transfer and cross-modal feature re-projection, efficiently extracts features related to the HER2 state from a conventional hematoxylin-eosin (H&E) staining section, and can directly accurately judge the HER2 state. Through the prior knowledge transfer and cross-modal feature re-projection mechanism, the application breaks through the bottlenecks such as insufficient utilization of prior knowledge and modal heterogeneity interference in the traditional method, and significantly improves the accuracy and robustness of the colorectal cancer HER2 state classification.

The invention discloses a mechanism knowledge base intelligent classification and indexing system and device based on multi-modal data fusion, and aims at solving the problems that when a traditional knowledge base processes multi-modal data, the efficiency is low, the classification precision is insufficient, the indexing quality is poor, and non-text resource processing is short. According to the system, through a feature extraction technology, features such as colors and textures are extracted for images, features such as keywords and themes are extracted for texts, and multi-type data such as texts, images and videos are converted into uniform feature vectors. A classification model is constructed by combining an SVM, a CNN and a Transform, and an attention mechanism is introduced to realize adaptive fusion. Based on a classification result, through automatic indexing such as keyword extraction and topic modeling, resources are associated through a knowledge graph. Experiments prove that the processing time of the device is shortened to 1 / 20 of that of an original method, the classification accuracy is improved from 30% to 90% or above, the knowledge base management efficiency and service quality are effectively improved, and support is provided for knowledge utilization and decision making.

The invention discloses an operation and maintenance management method, system and device based on artificial intelligence, and a medium, mainly relates to the technical field of operation and maintenance management, and is used for solving the problems of insufficient correlation analysis, low knowledge utilization efficiency and lack of self-healing ability in an existing scheme. Comprising the following steps: when a preset alarm self-healing processing program runs abnormally, progressively decreasing a quality evaluation value of a preset maintenance document set corresponding to the preset alarm self-healing processing program according to a first preset numerical value, adding the target alarm item, the equipment node corresponding to the target alarm item, the maintenance capability level, the quality evaluation value corresponding to the target maintenance document set and the maintenance explanation information to a preset manual maintenance management interface, and deleting the target alarm item, the equipment node corresponding to the target alarm item, the maintenance capability level, the quality evaluation value corresponding to the target maintenance document set and the maintenance explanation information from a preset automatic operation and maintenance management interface; and when the target maintenance document set does not exist, directly adding the target alarm item and the corresponding equipment node to a preset manual maintenance management interface.

The invention belongs to the technical field of automatic driving, and relates to an information processing method and electronic equipment. And target scene information is collected and fused through the multi-mode visual sensor, so that high-precision and standardized environmental perception is realized. And extracting related knowledge from the automatic driving world model knowledge base by utilizing a similarity matching model, and intelligently generating a cue word template conforming to a large language model input format based on the target visual information and the target knowledge information, so as to realize efficient semantic fusion. The large language model further processes input and outputs accurate target sensing information, and comprehensively reflects positions, motion states and categories of surrounding objects. In combination with a preset control rule, a reasonable target control instruction is generated, and knowledge availability verification is performed through the intermediate matrix to ensure that the instruction is scientific and effective. According to the method, sensor data and knowledge base information are effectively integrated, automatic driving environment perception accuracy and control decision credibility are improved, and technical guarantee is provided for safe and reliable operation of intelligent driving.

The application relates to an aspect-level sentiment analysis method based on a neighborhood window and knowledge graph enhancement, and belongs to the technical field of sentiment analysis. The method aims to solve the problems of semantic disorder, insufficient utilization of external knowledge and ineffective fusion of semantic and syntactic information in existing methods. The method first encodes the input sentence and aspect word, and generates a preliminary semantic representation. Then, a syntactic dependency graph is constructed and a multi-copy random walk algorithm is used for enhancement, and a graph convolution network is used to extract syntactic information. Next, the neighborhood window mechanism and the opinion alignment module are used to capture aspect-related semantic information from different perspectives and generate multi-perspective sentiment representations. Finally, the syntactic information, semantic information and knowledge information are fused, and a classifier is used for sentiment polarity prediction. Experimental results show that the application can effectively improve the performance and precision of aspect-level sentiment analysis and enhance the adaptability of the model to different domain data.

A method and system for extracting and organizing knowledge from historically accumulated data using AI assistance includes the following steps: unified preprocessing of multi-source heterogeneous aviation maintenance data to output unified JSON structured data; inputting the structured data into a large language model fine-tuned from aviation maintenance corpus for knowledge triple extraction; constructing a multimodal interaction and query interface supporting structured queries, natural language question answering, historical case retrieval, and maintenance solution recommendations; and establishing an incremental update mechanism to update newly submitted repair reports and batch-uploaded documents to the knowledge graph after preprocessing, extraction, and manual review. This invention effectively solves the problems of difficult standardization of multi-source heterogeneous aviation maintenance data, low reliability of knowledge extraction from large language models, and limited query and interaction methods, achieving efficient knowledge utilization and intelligent knowledge services from historical maintenance data.

This invention relates to the field of engineering design technology, and in particular to a method and system for auxiliary design of nuclear power plant engineering modifications. A method for auxiliary design of nuclear power plant engineering modifications includes: acquiring multi-format design documents and structuring the content of these documents to obtain structured data with consistent format; establishing a hybrid engine using a first language model and preset rules; extracting nuclear power plant engineering documents using the hybrid engine to construct a tag library; segmenting the structured data and saving it in association with the tag library to construct a nuclear power knowledge base; training a second language model based on the nuclear power knowledge base to establish a domain-specific vertical model; acquiring design instructions and generating auxiliary information using the domain-specific vertical model. This invention significantly improves knowledge utilization by leveraging structured data and a nuclear power knowledge base; ensures professionalism and accuracy through the combination of a vertical domain model and the nuclear power knowledge base; and creates an intelligent auxiliary design process using design instructions, improving design efficiency and document quality.