282 results about "Inference machine" patented technology

The invention discloses an electronic medical record intelligent evaluation method based on complex quality control indexes, and relates to the field of medical information processing and artificial intelligence. The method comprises the steps that an original electronic medical record text is collected and preprocessed, and structured diagnosis and treatment information and an event sequence diagram are extracted; through prompt word chain construction and a semantic reasoning mechanism, a large language model is guided to intelligently evaluate complex quality control indexes in medical records. The complex quality control indexes comprise diagnosis basis sufficiency, treatment scheme rationality, key result and change record integrity, treatment measure integrity and causal relationship rationality. According to the method, technologies such as a medical knowledge graph, a graph neural network and semantic vector retrieval are utilized to realize external knowledge recall and causal reasoning support; a self-consistency reasoning mechanism, a self-reflection mechanism and a multi-model cross validation mechanism are introduced to improve the accuracy and credibility of an evaluation result; and finally, outputting a structured quality control report and a visual reasoning chain. According to the method, the intelligence and refinement level in a complex medical quality control task can be remarkably improved, high interpretability and practical value are achieved, and the method is suitable for application scenes such as hospital quality management, scientific research evaluation and medical document standardization.

The invention relates to a multi-modal data and RAG technology fused power design knowledge base construction method, and belongs to the technical field of power software development. The method comprises the following steps: carrying out collection and information extraction on multi-source heterogeneous original data; the method comprises the following steps of: constructing a multi-dimensional knowledge element structure containing parameters, specifications and case relationships by carrying out classification, specialized and precise processing and cross-modal association on data; based on a vector, graph and relational database mixed storage architecture, semantic vector efficient retrieval, knowledge graph relation management and business data synchronization are achieved respectively; and a dynamic optimization result is subjected to hybrid retrieval, a dual-drive reasoning mechanism outputs compliance conclusions and bases, and a retrieval enhancement generation service ensures that output contents conform to specifications. And systematic management and intelligent application of the electric power design knowledge are realized.

The invention discloses an intelligent substation safety measure checking method and system, and the method comprises the steps: obtaining a secondary system topological structure, equipment information and historical safety measure ticket data of a substation, and constructing a quaternary knowledge graph; according to the quaternary knowledge graph, extracting multi-modal features of the maintenance task and identifying the type of a maintenance scene by combining a memory guide reflection decision reasoning mechanism, optimizing a rule reasoning process through a distributed guide local search algorithm, and generating an optimal safety measure operation set for a specific maintenance scene; automatically identifying a correlation loop and determining a minimum safety isolation range through a memory guide decision reasoning mechanism, dynamically generating a minimum safety operation set according to the real-time state of the equipment, and adaptively adjusting operation steps; and an operation dependency relationship model is constructed in combination with unwrapping variational multi-graph representation learning and a distributed guide local search algorithm, and operation sequence compliance verification, risk level assessment and dynamic visual early warning are realized. According to the invention, accurate formulation, dynamic adjustment and risk early warning of safety measures of the intelligent substation are realized.

The invention discloses an agent collaborative optimization data center management system based on knowledge graph driving, and relates to the technical field of data center management. The system specifically comprises the following modules: a modeling and entity management module, a cross-regional global scheduling optimization module, an agent game negotiation optimization module, an agent trust management module, a game negotiation conflict identification module, a reasoning evidence management module and a cross-regional consistency verification module. By establishing a complete knowledge graph model, systematic modeling of resource attributes, constraints, historical decisions and strategy preferences is realized, a unified data basis is provided for negotiation among multiple agents, the problem of a suboptimal solution caused by information asymmetry is effectively solved, a hierarchical reasoning mechanism is adopted, and the probability of resource disruption is reduced. In combination with global-region-node three-level reasoning and multi-round game negotiation, recursive optimization from global to local is realized, the reasoning complexity is effectively reduced, and the negotiation efficiency is improved.

The invention belongs to the technical field of computer data security, and discloses an enterprise data encryption storage system based on strategy optimization. The system is composed of a strategy map construction and optimization module, a sensitivity identification and encryption strategy generation module, an encryption execution and ciphertext structure processing module, a strategy-driven storage scheduling module and an audit recording and feedback optimization module. According to the method, a strategy graph containing multiple types of nodes such as users, data, behaviors, risks and resources is constructed, the edge weight and the node confidence coefficient are dynamically adjusted through the behavior frequency, the access relation and historical audit information, a graph structure oriented to behavior contexts is formed, and on the basis, an encryption strategy label is generated in combination with a fuzzy semantic reasoning mechanism; a traditional static recognition mode based on field rules can be broken through, potential high-sensitivity data and abnormal behavior paths are recognized, and the generation accuracy and the self-adaptive capacity of a data encryption strategy are improved.

The invention provides a method and a system for intelligently generating and dynamically adjusting a public emergency plan, which are oriented to the field of public safety emergency. The method comprises the following steps: fusing multi-source heterogeneous data to construct an extensible domain knowledge graph, establishing a standardized emergency instruction library, and carrying out multi-dimensional information labeling; automatic extraction of disaster elements is realized based on an entity recognition model of deep learning; analyzing association rules among the emergency entities through a semantic relationship mining technology; key information such as a disaster chain and resource distribution is rapidly obtained by using a map reasoning mechanism; carrying out cross-department resource collaborative allocation by adopting a multi-objective optimization algorithm to generate an optimal disposal scheme; a high-dynamic adjustment mechanism is constructed, and an emergency plan is dynamically optimized based on situation evolution prediction and real-time monitoring data; and finally, automatic generation and versioning management of the plan are realized. Through multi-mechanism cooperation of knowledge modeling, intelligent element analysis, semantic reasoning, dynamic optimization and predictive adjustment, the emergency plan generation efficiency, situation adaptability and resource allocation rationality are remarkably improved, and support is provided for quick response and scientific decision under emergencies.

The invention relates to an equipment fault intelligent question-answering method based on large model enhancement, and the method comprises the following steps: obtaining steel production line data, and constructing a fault knowledge graph; user questions are obtained and classified through semantic analysis, and classification results comprise single-hop questions, short-hop questions and multi-hop questions; calling a corresponding reasoning mechanism according to the category of the user question, generating an answer according to the fault knowledge graph, and outputting equipment fault analysis in a natural language form through a retrieval enhancement generation technology; for a single-hop problem, core entities and relationships are matched based on an Aho-Corasick automaton, and answers are directly retrieved from the fault knowledge graph; for a short jump problem, deriving an answer from the fault knowledge graph by adopting a single-agent inference model; for a multi-hop problem, a multi-agent collaborative reasoning model is adopted to generate a multi-step reasoning path based on the fault knowledge graph. Compared with the prior art, intelligent question answering with high reliability and high readability can be efficiently realized in a complex industrial scene.

The invention discloses an AI low-code development component automatic generation method based on a knowledge graph, and relates to the technical field of AI low-code development, and the method comprises the following specific steps: constructing an AI development knowledge graph: constructing a knowledge graph containing a five-layer structure, collecting and preprocessing multi-source data, and extracting various entities and association relationships, so as to obtain an AI development knowledge graph; constructing an ontology structure, importing the ontology structure into a database, and forming a final knowledge graph through manual auditing and machine reasoning optimization; according to the method, the multi-dimensional AI development knowledge graph is constructed, user requirements are analyzed in combination with the optimized natural language processing model, the component configuration scheme is generated by introducing the multi-level reasoning mechanism, and finally the standardized low-code component is automatically generated, so that full-automatic conversion from the user requirements to the low-code component is realized, and the user experience is improved. The defect that an existing low-code platform depends on manual component configuration is overcome, the development period is greatly shortened, and the development cost and the technical threshold are reduced.

The invention relates to an intelligent optimization method for an industrial waste transportation path, and the method comprises the steps: firstly collecting multi-source and multi-dimensional input variables, such as vehicles, roads, orders and waste attributes, achieving the dynamic simulation of a whole transportation process through digital twin modeling, and carrying out the structural processing of data through combining with space and environment labels; when an abnormal event is monitored, an abnormal correction action is dynamically generated and optimized through a multi-agent cooperation and game reasoning mechanism, and an optimal scheduling decision aiming at different working conditions is realized; and continuously monitoring and executing feedback by the system, returning the actual deviation to the model and the knowledge base, and driving the self-learning and continuous evolution of the anomaly correction strategy.

The invention discloses a Beidou-based new energy electric vehicle battery remote state monitoring system, and relates to the technical field of battery monitoring, and the system comprises a multi-dimensional data flow construction unit; a digital twin layer generation unit; the causal situation map establishing unit is used for generating a joint embedding vector, embedding the symbol constraint rule set into a preset graph neural network architecture, constructing a neural symbol causal discovery engine, directionally adjusting the joint embedding vector in combination with an anti-fact reasoning mechanism, and establishing a causal situation map; the contribution degree of each type of degradation inducement to the service life of the battery is evaluated; and the service life prediction and report generation unit is used for constructing a prediction generator and a verifier, outputting a consensus residual service life prediction value in combination with the real-time vehicle operation data, and generating a battery state report. According to the invention, the comprehensiveness and the real-time performance of data acquisition are improved, and the intelligent level and the application efficiency of remote monitoring and management of the battery are improved.

The invention relates to a meteorological data set automatic construction method and system based on modal bridging, and aims to meet the multi-modal large model training requirement in the meteorological field, and the method and system realize automatic conversion from an original meteorological image to a structured expert reasoning text through deep fusion of image and text information. The method comprises five stages of data preprocessing, image-text semantic modeling, causal reasoning generation, consistency screening and parallel processing, key meteorological elements are extracted by using a multi-modal model, a chained thinking process is constructed through a language model with meteorological knowledge, chained reasoning annotation is realized, cross-modal semantic alignment and a multi-round reasoning mechanism are introduced, and a multi-modal semantic model is established. The method is advantaged in that high-quality samples are screened in combination with rules and models, automation, high consistency and good expansibility are realized, manual annotation cost is substantially reduced, and the method is suitable for large-scale meteorological reasoning multi-modal data set construction.

The apparatus may be a wireless device configured to estimate, for a first transmission in a first slot, a first channel associated with the first transmission, wherein the first transmission is associated with a first precoding, receive, in a second slot following the first slot, a second transmission associated with a second precoding, and estimate, based on the received second transmission and at least one of the received first transmission or the estimated first channel, a second channel associated with the second transmission.

The invention relates to the technical field of automatic driving, and discloses an unmanned vehicle navigation method based on multi-modal fusion, which comprises the following steps: S1, acquiring and preprocessing sensor data, acquiring multi-modal sensor data, and performing data alignment, filtering and normalization processing; s2, constructing an optimization framework of the fusion weight of the sensor based on a variational inference method, defining an optimization objective function by using the preprocessed data as input, and solving an optimal solution of the fusion weight through the variational inference method; and S3, on the basis of the variational reasoning optimization model, optimizing a sensor weight updating process by utilizing information geometric measurement. A cross-modal learning and reasoning mechanism is introduced, weight updating is accelerated by using an information geometric optimization method, so that the unmanned vehicle can quickly adapt to sudden environmental changes, and compared with a traditional neural network model needing a large amount of training data, the method does not need large-scale data set pre-training, can adaptively adjust decisions, and improves generalization ability.

The invention discloses a monitoring and early warning method and system based on laboratory operation, and the method comprises the steps: obtaining and processing laboratory equipment state data and personnel operation records, and constructing a dynamic asymmetric Bayesian network which reflects the personnel operation intention, the equipment state and the interference relation; outputting probability distribution from the equipment state to the personnel operation intention, monitoring whether the equipment state is abnormal or not, and updating the probability distribution; performing judgment and early warning on the updated probability distribution based on an ROC curve, and defining a loss function to dynamically adjust the dynamic asymmetric Bayesian network; according to the method, the dynamic asymmetric Bayesian network is constructed, and parameters are dynamically adjusted according to the real-time change of the equipment state, so that the model shows higher robustness in a dynamically changing laboratory environment; by establishing a backward inference mechanism of the dynamic asymmetric Bayesian network, the operation intention of the personnel is inferred from the equipment state, the perception level of potential risks is improved, and the defects of a traditional system in deep inference are overcome.

The invention provides an emotion metaphor recognition method and device based on knowledge extraction and co-evolution reasoning, and relates to the technical field of natural language processing.The method comprises the steps that a high-quality emotion metaphor corpus is constructed, and advanced feature vectorization technology is combined to conduct deep processing on text data; key features are accurately extracted by using a fuzzy multi-granularity knowledge extraction technology, and an optimized feature subset is formed, so that redundant information is effectively removed, and the accuracy of feature selection is improved; furthermore, dynamic recognition and labeling of the emotion metaphor are achieved through a matching network and a co-evolution reasoning mechanism, and metaphor labels and emotion categories can be output at the same time. The method aims at solving the problems that in the prior art, emotion metaphor recognition precision is insufficient, the reasoning ability is limited, and noise and redundant information in large-scale text data are difficult to process.

The invention provides an equipment fault diagnosis recommendation method based on a knowledge graph and a multi-dimensional relevance index, and the method comprises the steps: converting a fault text into triple structured data based on an equipment fault knowledge graph ontology model, and constructing an equipment fault knowledge graph; constructing a multi-dimensional correlation index of the fault mode, and converting explicit feedback in the historical fault diagnosis data into implicit feedback; obtaining a fault mode which is strongly correlated with the abnormal phenomenon from historical fault diagnosis data based on a multi-dimensional correlation index, and taking the fault mode as a seed fault mode; and taking the seed fault mode as a starting point, iteratively propagating in the fault knowledge graph to obtain a ripple set, calculating an association probability between an abnormal phenomenon and a candidate fault mode according to the ripple set, accumulating each order of response according to the association probability, and outputting a prediction definite diagnosis probability. According to the method, the diagnosis relation strength between the abnormal phenomenon and the fault mode is revealed according to the multi-dimensional relevance index, and a flexible reasoning mechanism is adopted on the basis, so that the diagnosis accuracy and adaptability are improved.

The invention discloses a dynamic traffic signal control method based on a large language model, belongs to the technical field of artificial intelligence and intelligent traffic control, and aims to solve the problems of poor phase duration flexibility and weak adaptability caused by the fact that most traditional traffic signal control methods are limited to single-stage traffic phase control. According to the invention, the real-time traffic condition is input to the large language model in the form of natural language, and more efficient and intelligent traffic signal control is realized by using the strong generalization ability and the human-like reasoning mechanism of the large language model. The invention provides an efficient fine tuning architecture comprising two stages, the first stage training model learns an answer normal form and a reasoning track of a large parameter quantity model, and the second stage training model promotes the system to effectively learn excellent strategies and keep away from poorer strategies. The method provided by the invention can provide a new thought and technical support for urban intelligent traffic management in the future.

The invention belongs to the technical field of wireless communication, and discloses an edge computing and cache enabling meta-universe system and an intelligent optimization method, and the technical scheme of the invention comprises the following core optimization objectives: firstly, through an intelligent content cache strategy, a user cache hit rate is maximized, and unnecessary data transmission is reduced; secondly, task unloading decisions are optimized, computing resources are reasonably distributed, and energy consumption of user terminals is reduced; and thirdly, dynamic intelligent allocation of computing resources is realized, and the overall resource utilization efficiency of the system is improved. The ADRL algorithm provided by the invention has the following unique advantages: 1, by introducing an active reasoning mechanism, the decision ability of the algorithm in an uncertain environment is enhanced; secondly, in combination with preference information of the intelligent agent, an optimization strategy is more targeted; and thirdly, comprehensive balance of a multi-dimensional optimization target of the system is realized, and powerful technical support is provided for efficient operation of the element universe network.

The invention discloses a three-dimensional physical field real-time prediction method and system based on geometric deep learning and physical constraint, and relates to the technical field of computer-aided engineering and artificial intelligence. The method comprises the following steps: directly extracting native boundary representation data (B-Rep) of a three-dimensional model from a computer aided design system; constructing a heterogeneous dual graph taking a parameterized curved surface as a graph node, skipping finite element grid division, and aggregating local and global topological features by using a graph neural network; in combination with a physical information driving mechanism, a partial differential equation (PDE) residual error is introduced as a loss function for constraint training, and generalization prediction of a novel geometric structure is realized; and finally, the physical field state quantity is predicted through direct regression and is rendered in real time. An incremental reasoning mechanism based on a local topology subgraph is adopted, millisecond-level physical field real-time feedback under design modification is achieved, and the method is suitable for scheme rapid screening and trend prediction in the initial stage of design.

The invention discloses a personnel examination answer sheet identification method based on multi-agent collaborative perception. The method comprises the steps of obtaining an answer sheet scanning image, constructing a microstructure correction engine through reinforcement learning to process an answer sheet, scoring a correction result in combination with an unsupervised quality evaluation system, and adjusting correction parameters through a dynamic feedback iterative optimization mechanism. Deploying sensing multi-agent cooperative work, combining graph structure modeling and a template-free reasoning mechanism of a dynamic topology knowledge base, and combining a contribution degree evaluation mechanism to optimize the agents; entering a recognition stage of filling intention collaborative perception, modeling a recognition decision into a Markov decision process, quantifying recognition values of different scenes through a hierarchical reward mechanism, constructing a scene memory playback pool to store filling cases, and combining a lightweight online strategy optimization mechanism to obtain a recognition result and generate an evaluation report. According to the method, high-precision and high-robustness recognition of the complex answer sheet is realized, and the recognition accuracy and the paper marking efficiency are improved.

The invention relates to an equipment predictive maintenance management method and system based on multi-feature fusion, belongs to the technical field of equipment health management, and is used for solving the problems that existing sensor data is easily disturbed and distorted, and a potential causal structure of an equipment degradation path is difficult to reveal due to lack of a comprehensive modeling framework. The method comprises the steps of collecting multi-source data in real time, extracting a causal contribution degree of the data to a fault to generate a causal significance feature value, generating a three-dimensional health feature vector through an anti-fact neural network model in combination with a physical failure model, fusing the two to obtain an equipment state risk feature matrix, inputting the model to output a risk score, and obtaining an equipment state risk result. And finally, dynamically adjusting the monitoring frequency and the maintenance level and iteratively optimizing the strategy. According to the method, data interference can be filtered out, multi-class reasoning mechanisms are deeply fused, the equipment degradation law is accurately revealed, and full-life-cycle self-adaptive maintenance is achieved.

The invention discloses a vehicle-mounted distributed intelligent cooperative reasoning method. The method comprises the following steps: acquiring natural language input of a user in a vehicle; analyzing the natural language input through the vehicle-mounted large model to obtain a user intention; generating a plurality of reasoning tasks according to the user intention, and calling a plurality of reasoning models based on the plurality of reasoning tasks; generating a plurality of preliminary reasoning results based on the plurality of reasoning tasks through the plurality of reasoning models; the plurality of reasoning tasks are in one-to-one correspondence with the plurality of reasoning models; and generating a target reasoning result based on the plurality of preliminary reasoning results by using the vehicle-mounted large model. According to the method, the results after independent reasoning of each module are integrated to the vehicle-mounted large model in a centralized manner, and deep fusion and final decision making are performed on multi-system data by means of the vehicle-mounted large model, so that a cross-module collaborative reasoning mechanism is constructed, the user operation complexity is remarkably reduced, and the system response efficiency is effectively improved.

The invention belongs to the technical field of contract review, and particularly relates to an intelligent contract review analysis method and system based on a large language model. According to the method, dynamic generation of rules and a neural symbol reasoning mechanism are mainly fused, and the core is to realize automatic and precise risk identification and evaluation of contracts by utilizing the powerful language understanding and generation capability of a large language model (LLM); the method comprises the specific steps of generating a contract review rule, extracting facts from a contract to be reviewed, performing review reasoning and generating a review report. According to the method, the inherent logic preciseness of symbol deduction and the powerful semantic understanding capability of a large language model are organically combined, so that core elements and potential risks in a contract can be more accurately identified, and contract terms which are complex in expression or have hidden agreement can be effectively processed and expressed.

The invention belongs to the technical field of power transmission channel hidden danger recognition, and particularly relates to a power transmission channel hidden danger intelligent judgment method and device based on knowledge graph and multi-modal large model fusion. The method comprises the following steps: constructing a knowledge graph based on a power industry standard and a historical disposal record; acquiring a detection result of the hidden danger position of the power transmission site by using a target detection algorithm, mapping the detection result to the knowledge graph, and performing context verification and hidden danger level preliminary judgment based on a multi-level reasoning mechanism to obtain a corresponding graph reasoning result; a multi-modal large model is utilized to fuse visual information, a map reasoning result and a historical case, and a comprehensive evaluation report with professional logic is generated, so that automatic filtering of false alarms and accurate judgment of hidden danger levels are realized. The key defects in hidden danger identification and evaluation of the power transmission channel in the prior art are overcome.

The invention relates to the technical field of semantic understanding, in particular to a three-dimensional scene semantic understanding method and system based on multi-modal deep learning, and the method comprises the following steps: collecting a point cloud image and a depth map in an automatic driving scene, carrying out the normalization standardization and deletion filling, extracting texture geometric space features, and carrying out the fusion through an attention mechanism; a multi-time-step state vector is introduced to calculate change features, a spatial relation between road participation objects is modeled, a dynamic instance graph structure is constructed, semantic tags are reasoned, and fusion features are compared to generate a three-dimensional scene semantic understanding result. According to the method, the fusion quality is guaranteed through multi-source data normalization standardization, the semantic complementarity is enhanced through collaborative extraction of image texture and point cloud geometric features, the dynamic scene perception ability is improved through state vector modeling, the object interaction semantic relation is described through a spatial relation graph, and the recognition accuracy and consistency are improved through a semantic label reasoning mechanism. And the integrity and robustness of three-dimensional semantic understanding are integrally enhanced.

The invention relates to the technical field of text data processing, and discloses a text data extraction method, system, equipment and medium based on multi-modal fusion and self-evolution learning, which comprises the following steps of: performing feature extraction and spatial alignment on a printed text, a handwritten annotation and a dynamic table of a mixed format document to obtain a semantic feature of an image-text table, and inputting the semantic feature into a dynamic analysis layer; analyzing metaphor expressions and synonymous heterogeneous fields through field extraction and a context semantic reasoning mechanism, and outputting structured data; performing grammar compliance verification by adopting a regularization engine, and performing comparison verification through a federal learning mechanism; and inputting the verified data into the reinforcement learning model, updating the analysis rule and the model parameters through strategy iteration, and feeding back the updated analysis rule and model parameters to the dynamic analysis layer to complete closed-loop optimization. According to the method, the processing precision and efficiency of the complex document are greatly improved, the manual intervention requirement is remarkably reduced, and meanwhile, the privacy protection and compliance requirements are met.

The apparatus may be a wireless device configured to estimate, for a first transmission in a first slot, a first channel associated with the first transmission, wherein the first transmission is associated with a first precoding, receive, in a second slot following the first slot, a second transmission associated with a second precoding, and estimate, based on the received second transmission and at least one of the received first transmission or the estimated first channel, a second channel associated with the second transmission.

The invention discloses a monitoring data trend analysis and decision suggestion generation method based on a large language model agent. The method comprises the following specific steps: step 1, constructing a domain knowledge enhancement module; 2, developing and registering a callable tool set, and compiling a series of performance functions by using Python; 3, constructing an agent arrangement platform and a core agent, and creating a master control agent in the agent arrangement platform as a scheduling center of an analysis task; step 4, executing an intelligent agent analysis process, and when an analysis request is received, enabling the intelligent agent to autonomously run according to the following process; 5, system integration and interface calling are carried out, the whole agent is packaged into RESTful API service and deployed in an enterprise server, by introducing a large language model agent architecture, an RAG enhanced reasoning mechanism and a callable tool set, the technical transition from passive data display to active insight generation is achieved, and the method has the advantages of being high in practicability and easy to popularize. And the analysis efficiency, the decision accuracy and the system intelligence level are remarkably improved.

The invention discloses a text risk identification method and device fusing a fine-grained knowledge graph and deep learning, and the method comprises the steps: carrying out the combined entity identification of a standardized text set, and extracting evaluation dimensions and viewpoint contents to form an entity pair set; performing sentiment polarity analysis on the entity pair set to obtain sentiment tendency, and constructing a triple (evaluation dimension, sentiment tendency and viewpoint content); constructing a fine-grained emotional knowledge graph based on the triple, and optimizing the fine-grained emotional knowledge graph by adopting structure perception and joint loss; based on the to-be-recognized text, retrieving from the atlas to construct an enhanced text, and inputting a fusion atlas perception reasoning model comprising a BERT coding layer, an atlas attention module, a cross-layer residual attention mechanism and a KAN reasoning layer to obtain a text risk recognition classification result. According to the method, an emotion structure modeling and semantic reasoning mechanism is introduced, the accuracy, interpretability and robustness of text risk recognition are improved, and the method is suitable for multi-field text analysis tasks under complex contexts.

The invention relates to the technical field of information security, in particular to a sensitive information detection technology and system based on multi-mode and steganography detection.The sensitive information detection technology and system based on multi-mode and steganography detection.The method comprises the steps that firstly, through a format tampering prevention module, file header information and binary data analysis are utilized, the real type of a file is verified, and format camouflage attacks are prevented; the sensitive data detection module adopts DistilBERT and LSTM structures to preprocess a text, extract text embedding, mine text features through a bidirectional LSTM and Attention mechanism, and finally accurately judge whether the text contains sensitive information or not; the joint content and file steganography detection module fuses a convolutional neural network and multi-modal learning, adopts corresponding models for different data types to carry out steganography information detection, integrates results through a multi-modal neural network, and applies multi-modal feedback and a cross-modal reasoning mechanism to improve the detection accuracy; according to the method, the accuracy, comprehensiveness and system adaptability of sensitive information detection are effectively enhanced.