1050 results about "Pattern matching" patented technology

The invention relates to the technical field of image processing, and discloses an intelligent machine vision detection method and system based on image processing, and a storage medium. The method comprises the steps of performing candidate region generation and geometric constraint optimization processing on fusion features through a cascade bidirectional region suggestion network to obtain a target candidate region, and performing instance segmentation processing on the target candidate region through a graph convolutional network and a boundary dynamic adjustment mechanism to obtain a segmentation result; performing feature extraction processing on the segmentation result through a space-time joint attention network to obtain space-time context features; performing multi-target tracking processing on the spatio-temporal context features through depth correlation measurement and motion mode matching to obtain a trajectory prediction result; and performing abnormal mode modeling processing on the trajectory prediction result through hierarchical contrast learning and multi-modal feature fusion to obtain abnormal features, and generating a visual inspection report according to the abnormal features. According to the invention, the accuracy of intelligent machine vision detection based on image processing is improved.

The invention provides a supply chain multi-dimensional fluctuation prediction and inventory optimization method based on artificial intelligence, and relates to the technical field of supply chain management, and the method comprises the steps: obtaining multi-dimensional data, carrying out the preprocessing of the multi-dimensional data, generating a standardized time series data set and a knowledge graph, carrying out the multi-scale decomposition through wavelet transformation, and extracting fluctuation features; and constructing a state matrix by combining a graph attention network and a multi-head self-attention mechanism, predicting a risk conduction path based on historical fluctuation pattern matching and a graph convolutional network, and generating a risk prevention and control scheme. According to the invention, the fluctuation characteristics and the risk conduction path in the supply chain system can be accurately identified, and the early warning accuracy and the response timeliness are improved.

The invention discloses a network threat multi-modal detection method based on a large model, and belongs to the technical field of network security, and the method comprises the steps: collecting three types of heterogeneous data of NetFlow flow of a network layer, a system call chain sequence of a host layer and a protocol load of an application layer, and carrying out the desensitization processing and feature coding to generate a unified tensor format; the method comprises the following steps: through network security threat intelligence and MITRE ATTamp; performing supervision fine tuning on the large model by using a CK attack chain sample, and constructing a network threat identification special model; cross-device behavior characteristics are extracted through a model self-attention mechanism, and a dynamic behavior map is constructed; and finally, comprehensively evaluating an attack mode matching degree, a node vulnerability mean value, historical alarm association and an attack path risk by adopting a weighted fusion algorithm, and triggering a high-confidence alarm when a comprehensive score exceeds 0.8. According to the method, through multi-modal data fusion and dynamic graph analysis, the detection precision and response efficiency of the complex attack chain are improved.

The invention relates to the technical field of image processing, and discloses a video monitoring personnel behavior identification method and system. The method comprises the following steps: performing action node and joint angle extraction processing on a positioning area to obtain human body action skeleton sequence data; constructing an action connection graph and a time sequence association edge, and carrying out behavior pattern learning processing to obtain space-time behavior characteristics; performing action semantic mapping processing on the spatio-temporal behavior features and the human body behavior feature vectors, and performing behavior time sequence chain construction processing to obtain behavior semantic description features; and performing behavior pattern matching and anomaly measurement processing on the behavior semantic description features, performing feature discrimination processing to obtain a behavior type identifier and an anomaly degree index, performing behavior persistence analysis processing, and performing behavior risk level evaluation processing to obtain a behavior recognition result and early warning information. According to the invention, based on multi-modal feature fusion and spatio-temporal behavior modeling, accurate identification of personnel behaviors and dynamic assessment of risk levels are realized.

The invention provides a photovoltaic active distribution network voltage out-of-limit problem tracing method comprising the following steps: receiving research data and carrying out multi-scale feature analysis to obtain a standardized multi-dimensional feature matrix and a time-space incidence matrix; constructing an initial causal graph based on basic research data and power grid physical constraints, and constructing a multi-layer causal network by calculating causal intensity; extracting a dynamic feature sequence, analyzing propagation features, and generating a propagation path set and a path weight matrix; performing tensor completion and state estimation on the research data to obtain a state estimation matrix and estimation uncertainty; and outputting an out-of-limit source position, an out-of-limit risk score and early warning information based on pattern matching analysis and risk assessment. According to the invention, through combination of multi-dimensional feature analysis and a multi-layer causal network, accurate positioning and risk early warning of a voltage out-of-limit source are realized, and traceability accuracy and analysis efficiency are improved.

The invention relates to the technical field of data exchange, in particular to a data exchange system of a nested tag structure, which comprises a data carrier adaptation module, a structure analysis module, a field mapping module, a logic driving module and a script generation module. According to the method, carrier identifiers are generated through comparison of MIME types, file header features are positioned through pattern matching, the recognition accuracy and adaptability are improved, a nesting relation is calculated through depth-first traversal of file paths, node sequences are generated through analysis in combination with a DOM tree, the label analysis efficiency is optimized, and a mapping table is generated through algorithm matching of field identifiers and database codes; the method comprises the following steps: dynamically constructing a field mapping relation, reducing manual configuration, constructing a decision matrix based on a mapping table, executing logical operation to generate a status bit and triggering a threshold write-in instruction, enhancing strategy dynamic adjustment capability, reducing semantic ambiguity through multi-algorithm collaboration and a dynamic decision mechanism, improving data exchange robustness, and supporting real-time state response. And the transmission timeliness of the industrial Internet of Things is ensured.

The invention provides an electric power Internet of Things security response method and system based on risk assessment, and the method comprises the steps: obtaining a heterogeneous security data set, and carrying out the data processing of the heterogeneous security data set based on a preset data processing rule, and obtaining a standardized data set; according to ATTamp; according to a CK causal relationship analysis method, pattern matching is carried out on the standardized data set, and whether a potential threat scene exists or not is recognized; if the potential threat scene exists, performing time sequence correlation analysis on the threat scene according to a preset correlation graph to obtain a correlation analysis result; performing link analysis on the correlation analysis result to obtain a key attack path; inputting the key attack path and the Internet of Things terminal access behavior data in the standardized data set into a preset abnormal behavior classification model, so that the abnormal behavior classification model judges whether an abnormal behavior exists or not; if it is judged that the abnormal behavior exists, safety response is carried out based on preset response measures, and the operation safety of the Internet of Things terminal is improved.

The invention discloses a document information input method and system based on big data processing, and belongs to the technical field of document information processing. According to the method, deep risk feature screening is performed on document data by combining big data mining and a knowledge graph, potential sensitive elements can be accurately identified, risk categories and influence ranges are analyzed based on association rules, automatic early warning and real-time risk notification are realized, a dynamic rule knowledge base is adopted, multi-dimensional compliance verification is supported, and the risk assessment efficiency is improved. Service constraint conditions can be updated in real time, the flexibility and adaptability of rules are ensured, the accuracy and timeliness of compliance check are improved, pattern matching and multi-level check are performed on compliance data objects through an intelligent check process and a recheck rule set trained by big data, the preciseness and reliability of data check are ensured, and the data check efficiency is improved. Through comprehensive quantitative calculation of the business complexity, the risk coefficient and the timeliness priority, the task scheduling weight parameter is generated, the receipt auditing path is optimized, and the auditing efficiency is improved.

The invention provides a monitoring information identification method based on the Internet of Things and a server, and the method comprises the steps: firstly obtaining multi-modal perception data which are collected by a plurality of Internet of Things monitoring nodes in a target monitoring area in real time and comprise images, sounds and environment parameters, and then carrying out the cross-modal feature extraction and fusion processing of the multi-modal perception data; the method comprises the steps of generating a comprehensive spatio-temporal characteristic spectrum, determining abnormal event types and spatio-temporal distribution information in a region through dynamic mode matching based on a preset abnormal mode library, generating an adaptive regulation and control instruction set according to the abnormal event types and the spatio-temporal distribution information, and determining control parameters and execution priorities of different Internet of Things execution devices. And finally, distributing the instruction set to an associated Internet of Things execution device cluster, triggering the cluster to execute control parameters according to priorities, eliminating the influence of abnormal events, and realizing comprehensive and intelligent monitoring and exception handling of a target monitoring area.

The invention relates to the technical field of AI protection, and discloses an AI protection engine construction method and system based on Web application. The method comprises the steps of obtaining a Web application HTTP request and performing semantic analysis to obtain request feature data; performing multi-dimensional threat feature extraction to obtain a Web request threat feature set; performing mode matching on the Web request threat feature set and a preset attack mode library to obtain a threat type judgment result and attack intention information; performing time sequence behavior analysis to obtain a Web behavior abnormal index; comprehensive decision making is carried out through a multi-agent collaborative decision making system, and a defense decision making scheme is obtained; according to the defense decision scheme, a corresponding defense component is selected through a defense execution engine for safety protection, and a Web request processing result is obtained. According to the method, corresponding protection measures can be taken aiming at requests of different risk levels, so that logic vulnerability attacks which are difficult to detect by a traditional system are effectively identified and defended.

The invention discloses a fault root cause positioning method and system driven by a dynamic knowledge graph, and relates to the technical field of fault root cause localization, and the method comprises the steps: collecting and obtaining a multi-source fault associated data set, carrying out the entity association extraction of the multi-source fault associated data set, and obtaining a fault entity set and an entity relationship set; performing graph node cascading and incremental learning updating, and constructing a fault updating knowledge graph; monitoring and acquiring target fault data, performing mode matching reasoning, and generating a fault mode candidate root cause set; and performing similarity matching on the fault mode candidate root cause set in combination with a historical fault case library, and determining a target fault root cause positioning result. The technical problem of low fault diagnosis efficiency caused by inaccurate fault root cause positioning and knowledge graph updating lagging in the prior art is solved, and the technical effects of realizing accurate positioning of the fault root cause and dynamic improvement of the knowledge graph and improving the fault diagnosis efficiency and accuracy are achieved.

The invention discloses a moving target trajectory prediction method and related equipment, and relates to the technical field of trajectory prediction.The method comprises the steps that multi-source dynamic sensing data of a moving target are obtained, and the multi-source dynamic sensing data comprise target motion state parameters, environment disturbance parameters and biological behavior characteristic parameters; generating an initial trajectory prediction result based on the target motion state parameters; dynamically correcting the initial trajectory prediction result based on the environmental disturbance parameters to obtain a disturbance compensation trajectory; and performing behavior pattern matching on the disturbance compensation trajectory based on the biological behavior characteristic parameters to generate a target trajectory prediction result. By fusing the multi-source dynamic sensing data and combining a dual optimization mechanism of environment disturbance dynamic correction and biological behavior pattern matching, the trajectory prediction precision in a complex scene is improved.

The invention relates to the technical field of data security analysis, in particular to a data security analysis system and method based on artificial intelligence. The method comprises the following steps: a data processing unit performs parallel cleaning and standardization processing on network flow data, user behavior data and system log data, and integrates the data into a comprehensive feature vector; the anomaly detection unit uses a DBSCAN algorithm to carry out clustering analysis on the comprehensive feature vectors, and a Gaussian mixture model is used to carry out probability modeling to screen abnormal points; the behavior analysis unit determines the type of an abnormal point through mode matching, carries out risk assessment on the abnormal point, and generates a safety analysis report; and the safety response unit adopts an automatic response mechanism based on the abnormal point risk level and generates a safety response report. According to the data security analysis system and method based on artificial intelligence, efficient abnormal point division is realized by dynamically adjusting clustering parameters and a density sensing technology, and abnormal detection and security risk assessment are realized by combining a Gaussian mixture model and a posterior probability screening mechanism.

The invention provides a network attack detection method and system based on a distributed intelligent probe, and the method comprises the steps: receiving real-time flow data synchronously collected by the distributed intelligent probe at each node of a network, carrying out the inter-node interaction relation modeling processing of the real-time flow data, recognizing a flow communication mode between different network nodes, and carrying out the detection of the network attack. Generating a flow association map containing the node connection relationship and the communication frequency; carrying out abnormal communication path mining based on the flow association map, and extracting a node communication sequence with an abnormal mode by analyzing the deviation degree of a node connection relationship and the fluctuation characteristics of communication frequency; performing pattern matching processing on the node communication sequence and an attack behavior template in a preset attack feature library, calculating sequence matching similarity and generating an attack matching degree score set; and determining a network attack type and attack source node positioning information, and generating a network attack detection result. According to the invention, the practicability and effectiveness of network attack detection are improved.

The invention provides a pattern multi-row matching method and device for ATE equipment, and the method comprises the steps: determining a cyclic pattern sequence comprising a plurality of rows of target patterns based on the number of matched rows when a first to-be-tested pattern carrying a LOOPMATCH instruction is recognized after a test is started; lOOPMATCH circulation is executed on the basis of the circulation Pattern sequence, and in each LOOPMATCH circulation, the Pattern matching result of the target Pattern in each row is determined; when all Pattern matching results pass in one cycle, jumping out of the LOOPMATCH cycle and determining that the cycle Pattern sequence test passes, otherwise, continuing to circulate to the maximum cycle number and then determining that the cycle Pattern sequence test does not pass; according to the technical scheme of the embodiment of the invention, when the first to-be-tested Pattern carrying the LOOPMATCH instruction is detected, multiple rows of target Patterns are constructed for loop testing, and LOOPMATCH loop is jumped out after any complete matching succeeds, so that the depth of the Pattern sequence can be effectively reduced, multi-row synchronous multi-time level matching is realized, the test resource overhead is reduced, and the test efficiency is improved.

The embodiment of the invention provides a fraud phone real-time recognition method and device based on AI semantic understanding, and the method and device achieve the precise understanding of the dialogue content through the innovative construction of a voice analysis mechanism, the grammatical feature extraction and the semantic role marking. And designing a scene discrimination model based on verbal skill recognition, and establishing a fraud verbal skill recognition strategy for intelligent classification in combination with semantic pattern matching and hierarchical clustering algorithms. A residual fusion assessment mechanism is introduced, and accurate assessment and timely prevention and control of call risks are realized through historical case feature fusion and risk scoring. According to the method, the defects of the traditional technology in the aspects of speech understanding, verbal skill recognition, risk assessment and the like are effectively overcome, and the accuracy and reliability of fraud phone recognition are remarkably improved.

The invention relates to the technical field of data processing, and discloses a multi-source heterogeneous big data processing system, which comprises the steps of dividing data into structural sub-data and non-structural sub-data, performing feature extraction on the non-structural sub-data to obtain multi-modal features, so that the data can be compared in a unified feature space, and then, performing feature extraction on the non-structural sub-data to obtain multi-modal features. The similarity and relevance between the data are identified through pattern matching, it is ensured that the data of the same entity are identified to be directly fused, redundancy is reduced, accuracy is improved, then the most similar data combination is selected through an optimization algorithm, and the actual relation type between the data is determined through relation deduction; and finally, the most suitable fusion mode is selected according to the relationships. According to the method, the accuracy and effectiveness of the data fusion process are ensured, and a reliable basis is provided for deeper analysis and decision making.

The invention provides a processing method and system for reconstructing an LDI sub-pixel image through sparse compression and application, and the method comprises the steps: calculating a low-resolution to-be-exposed image corresponding to each phase structure in advance through a computer image compression algorithm; the low-resolution to-be-exposed images obtained through calculation are loaded to the digital micro-reflector according to the time sequence, and the display moment of each pattern is matched with the rotation position of the corresponding phase structure; coupling the pattern of the digital micro-mirror and the phase encoding mask in a frequency domain through a 4-f optical system, and reconstructing a high-resolution exposure pattern at a sub-pixel level by utilizing a diffraction effect; projecting a high-resolution light field of the reconstructed high-resolution exposure pattern to the surface of the photoresist, and forming a target circuit pattern by accumulating exposure dose; and the micro-nano structure with sub-pixel precision is obtained after development. The system comprises an image compression module, a pattern matching module and a pattern forming module. According to the invention, the manufacturing cost of the laser direct imaging equipment is reduced under the condition of the same precision.

A system may receive, via a cloud-based platform supporting a plurality of tenants, one or more access events from a user of a host platform associated with a tenant, the one or more access events comprising one or more keystroke events at the host platform and one or more commands inputted by the user at the host platform, wherein the one or more access events are captured by a continuous authentication agent associated with the host platform. The system may identify fraudulent access events based at least in part on executing a machine learning model to perform a pattern matching between previously authenticated browsing behavior of the user and the one or more access events at the host platform. The system may generate a challenge question for reauthenticating the user of the host platform based at least in part on identifying the at least one fraudulent access event.

The invention discloses a low-voltage transformer area-oriented power failure fault research and positioning method and system, and relates to the technical field of electric power, and the method comprises the steps: collecting power failure event data of a low-voltage transformer area in real time; performing multi-dimensional feature analysis on the power failure event data, and constructing a transformer area topological relation network according to space-time correlation features; generating a fault feature vector based on the space-time correlation feature and the transformer area topological relation network, performing mode matching according to the fault feature vector, and determining a fault matching result; and carrying out positioning verification on the low-voltage transformer area according to a fault matching result, and generating a fault positioning parameter. Therefore, the technical effects of quickly and efficiently realizing the power failure fault diagnosis and positioning in the low-voltage transformer area and improving the positioning accuracy and reliability are achieved.

The invention provides a full-layout defect rapid detection system based on pattern matching and classification, which relates to the technical field of semiconductor manufacturing and comprises a data acquisition and multi-source fusion module, a self-adaptive preprocessing module, a dynamic pattern matching module, a multi-source feature fusion module and a closed-loop optimization and output module, the dynamic mode matching module dynamically generates a defect template from real-time data through a clustering algorithm (such as DBSCAN) and a generative adversarial network (GAN), the speed and accuracy of defect detection are remarkably improved in combination with reinforcement learning filtering and region focusing technologies of the self-adaptive preprocessing module, a detection strategy can be adaptively adjusted according to the real-time data, and the defect detection accuracy is improved. According to the method, a high-risk area is preferentially matched, and GPU parallel computing acceleration processing is performed, so that the detection time is shortened to 50% or below of that of a traditional method, deformation or fuzzy defects can be effectively identified, the false alarm rate is reduced by about 30%, and high-quality input is provided for subsequent matching and classification.

The invention provides an ecological circulation barrel-in-barrel culture data acquisition method and system, and relates to the technical field of intelligent control, and the method comprises the steps: carrying out the feature extraction of a historical data flow through a multivariable coupling analysis model, carrying out the mode matching through a preset environment threshold interval and a culture environment state matrix, and obtaining a culture environment state matrix; generating a water quality health degree evaluation index and an ecological imbalance early warning signal; and analyzing the water quality health degree evaluation index and the ecological imbalance early warning signal based on a fuzzy control algorithm, dynamically calculating a final feeding amount curve and a water change intensity function, and generating a regulation and control instruction set containing the feeding frequency, the bait particle size and the water pump rotating speed. According to the invention, real-time evaluation and accurate intervention of the water quality health state can be realized.

The invention relates to the field of energy consumption analysis, and discloses an AI-based energy consumption data analysis and prediction system, which comprises the steps of collecting environmental parameters and running states of equipment, dynamically identifying the current system working condition by using a working condition identification algorithm combining incremental clustering and historical mode matching, and predicting the energy consumption data. Collected data is divided according to time, space and working condition dimensions, multi-scale features are extracted, normalization parameters can be dynamically adjusted along with changes of working conditions, a drift index is calculated through comparison of a drift threshold value and historical distribution, an optimal normalization updating strategy is selected according to the drift index, the normalization parameters are dynamically updated, and energy consumption trend prediction is conducted through a statistical model. A prediction result is combined with a working condition label to carry out weighted correction, error analysis and deviation detection are carried out in combination with a drift index, working condition prediction and historical error data, and an analysis result is fed back to a working condition sensing module, a feature adaptive module and a normalization control module. The method has the advantage of improving the stability and reliability in a dynamic environment.

The invention discloses a government affair data automatic classification and grading method, and relates to the technical field of data management and information processing. According to the method, data of different sources and formats are converted into structured formats by adopting a regular expression and a pattern matching technology, and the consistency and accuracy of the data are improved by means of a machine learning algorithm, data quality detection, anomaly correction and the like; by constructing a government domain ontology and a knowledge graph and combining a natural language processing technology, semantic classification is effectively carried out on government affair data, overlapping and uncertainty between categories are solved, robustness and accuracy of classification decision are further improved through an integrated learning algorithm, high efficiency and reliability of classification are ensured, and the method is suitable for large-scale popularization and application. By establishing a data grading dynamic adjustment and update mechanism and combining real-time data monitoring, sensitivity evaluation and automatic adjustment of access control rules, the flexibility and security of data management are ensured.

The invention relates to the technical field of equipment fault prediction, and discloses a fire fighting equipment fault early warning method and system based on multi-dimensional data linkage, and the method comprises the steps: collecting multi-dimensional operation parameters, and carrying out the data preprocessing, and obtaining standardized data; and then time synchronization and dependency relationship analysis are carried out, an equipment dependency graph is constructed, influence weights among equipment are quantified, and abnormal fluctuation characteristics are extracted to form a risk mode set. And based on the risk mode set, identifying a high-frequency high-weight risk cluster as a preliminary fault feature through grouping analysis, and determining a core fault feature by combining historical data correlation analysis. And then matching abnormal signals to generate spatial labels, integrating data to form an abnormal event candidate list, analyzing an event propagation path in combination with environmental data, and verifying the event propagation path to obtain a final abnormal event list. And finally, performing priority grading according to a historical mode matching result, and generating a visual early warning report by using a three-dimensional point cloud rendering technology. According to the method, the fault early warning accuracy of the fire-fighting equipment is improved.

The invention provides a cross-protocol identification data intelligent analysis middleware method and system. The method comprises the following steps: taking multiple industrial communication protocol identifiers in advance to form a metadata set, and constructing a protocol feature element model library containing a non-standard protocol feature template; a programmable logic device is introduced, physical layer signal waveform characteristics are matched, and corresponding communication protocol stacks are dynamically switched at a hardware level according to a matching result; receiving a communication data stream through an electric domain processing unit of the programmable logic device, and generating a protocol analysis logic sequence in combination with a state transition rule of the meta-model library and a mode matching rule of the protocol stack; performing combined operation on the structured feature template and the analysis logic sequence to generate an executable cross-protocol data analysis rule; and when unknown protocol characteristics are detected, triggering dynamic reconstruction of the template, so that the analysis rule is adapted to a newly added protocol. According to the technical scheme provided by the invention, the protocol compatibility development cost in an industrial multi-protocol scene is remarkably reduced.

The invention relates to the technical field of intelligent power grid and household energy management, and provides a household ultra-short-term load prediction method, which comprises the following steps of: performing preprocessing and characteristic engineering processing on user power consumption data acquired by an intelligent electric meter; user type clusters are obtained through clustering analysis, and a standard user power consumption mode corresponding to each cluster is determined; constructing a deep learning prediction model based on the mode, and generating a model parameter matrix sequence and a timestamp; generating an ultra-short-term load prediction value by using the parameter matrix and historical data, and forming a prediction load curve according to a timestamp; and finally, the prediction curve and the management suggestion are sent to the household energy management terminal. According to the method, accurate prediction is realized through data cleaning, user grouping, pattern matching and deep learning modeling, and a visual result and an energy-saving suggestion are output. According to the invention, the household energy management efficiency and the power utilization economy can be improved.

The invention discloses an automatic monitoring method for infrastructure construction operation violation based on multi-modal data fusion, and belongs to the technical field of graphic data processing and pattern recognition, and the method comprises the steps: obtaining multi-modal source data of a target scene, carrying out the data recognition and standardization processing of the multi-modal source data, and generating preprocessing feature data; acquiring an on-site environment condition, and generating an environment state parameter; performing dynamic weight distribution on the preprocessed feature data based on the environment state parameters to generate weighted feature data; performing multi-source data conflict judgment and resolution on the weighted feature data to generate conflict resolution event data; and based on a preset behavior rule pattern library, performing behavior pattern matching on the conflict resolution event data to generate a violation judgment result. According to the method, a multi-modal spatial-temporal feature fusion and dynamic weight distribution mechanism is adopted, and an environment adaptive perception and multi-source data conflict resolution technology is combined, so that automatic monitoring and pattern recognition of illegal behaviors in a construction environment can be realized.

The invention relates to the technical field of database security, and discloses a semantic analysis-based SQL (Structured Query Language) risk check method and device, and the method comprises the following steps: analyzing based on an input SQL statement to obtain a standardized SQL statement; constructing an abstract syntax tree according to the standardized SQL statement; extracting key semantic information according to the abstract syntax tree; obtaining a risk assessment result according to the key semantic information; and recording an audit log according to the risk assessment result. Compared with the technical problem that a traditional auditing system mainly depends on pattern matching of basic rules, and potential safety risks cannot be accurately recognized when complex SQL statements are processed, especially under the condition that multi-layer nesting and dynamic construction are included, due to the fact that the efficient semantic analysis technology and the risk assessment algorithm are combined for use, the safety risk assessment method and device have the advantages that the safety risk assessment efficiency is improved, and the safety risk assessment efficiency is improved. Therefore, the occurrence of security events is effectively avoided, the accuracy and efficiency of SQL risk check are improved, and then the security of the database is ensured.

The invention provides a large-scale language model knowledge editing method and device based on pattern matching, and the method comprises the steps: obtaining target editing knowledge which is used for carrying out the knowledge updating of a large-scale language model; determining a plurality of mode matching degrees between each level in the large language model and the target editing knowledge; determining the highest mode matching degree in the plurality of mode matching degrees, and taking the hierarchy corresponding to the highest mode matching degree as a target editing layer; performing rank-one editing on the multi-layer perceptron of the target editing layer based on the target editing knowledge to obtain an edited large language model; according to the method and the device, other irrelevant data is not influenced while knowledge editing is accurately carried out.