96 results about "Source tracing" patented technology

The invention belongs to the technical field of river and lake water quality monitoring, and discloses a river and lake water quality intelligent traceability monitoring system and method.Dissolved organic matter fluorescent fingerprints are analyzed in real time through a three-dimensional fluorescent monitoring unit, pollution markers are detected, and parameters such as COD, ammonia nitrogen and total phosphorus are synchronously detected through a full-spectrum monitoring unit; the video monitoring unit is used for collecting section images / videos of rivers and lakes, other sensors are used for collecting real-time values of chlorophyll, pH, conductivity, turbidity and dissolved oxygen in a water body, then the data processing module is used for carrying out normalization processing on multi-source data, and the analysis and early warning module is used for analyzing water quality early warning scores, spectrum early warning scores and video early warning scores. All scores are synthesized to obtain a comprehensive early warning score, traceability is started when comprehensive early warning is larger than or equal to the grade II, pollution sources are recognized based on three-dimensional fluorescence spectrum comparison and upstream patrol, and a river and lake water body monitoring-early warning-traceability-verification closed loop is achieved.

The invention discloses an intelligent construction and tracing method and device for an attack graph, and relates to the technical field of network security. The method comprises the steps of performing semantic analysis and entity relationship extraction on a multi-source heterogeneous security log according to a predefined structured security data model, and generating a standardized security entity relationship triple set; based on the set, taking an entity in an initial alarm as a starting point, and adopting an iterative closed loop driven by a large language model to dynamically construct an attack graph; and carrying out attack technique and tactics mapping and threat attribution based on the final map, and generating a response strategy of priority ranking. According to the method, automatic and high-precision source tracing and response of the attack chain are realized, and the problems that the prior art depends on static rules and semantic segmentation and lacks dynamic reasoning capability are effectively solved.

An attack detection and source tracing method includes acquiring entities in a target network environment and interaction event information between the entities and constructing a network event graph with the entities and the interaction event information; determining a graph embedding vector of each interaction event information in the network event graph as feature information based on a preset attack feature recognition model and determining an attack event in the network event graph according to the feature information; determining dependencies between the attack event and remaining interaction event information in the network event graph and searching for a corresponding interaction event information as source tracing information of the attack event according to the dependencies.

The invention discloses a network threat evidence fixation technology for solving source tracing evidence chain deficiency, and belongs to the field of network security. Aiming at the problems of distributed log time sequence chaos, easy tampering of virtual environment metadata, difficult association of cross-platform attack traces and the like existing in the traditional evidence obtaining technology, the invention constructs a technical system of dynamic data capture and space-time fusion modeling, evidence chain reconstruction and distributed verification, anti-quantum storage and security solidification. A self-adaptive sensor network is constructed through an improved BFGS algorithm, a space-time fusion evidence model is provided, and an improved MPT structure and a lattice-based cryptographic algorithm are designed in combination with a time sequence Petr i network and a Byzantine fault-tolerant protocol. Tests show that the attack feature capture rate reaches 91%, the cross-platform attack association accuracy rate reaches 93%, the quantum signature resistance speed is 1200 times per second, and the problem of source tracing evidence chain deficiency is effectively solved.

The invention provides a rapid source tracing method and system for a pollution source based on dual unmanned aerial vehicle cooperative gradient detection, and belongs to the technical field of environment monitoring and unmanned aerial vehicle application, and the method comprises the steps: initializing system parameters; deploying the first unmanned aerial vehicle to hover at an initial position, and measuring an atmospheric pollutant concentration value at the initial position and recording the atmospheric pollutant concentration value as a first concentration value; the second unmanned aerial vehicle is controlled to fly around, and the pollutant concentration value is measured and recorded as a second concentration value; calculating a pollutant concentration gradient value; based on the concentration gradient value, the direction with the maximum atmospheric pollutant concentration gradient in the whole-cycle range of 0-360 degrees is determined and recorded as the maximum gradient direction; the second unmanned aerial vehicle is controlled to move to the intersection point position of the maximum gradient direction and the fly-around circumference for hovering; the first unmanned aerial vehicle is controlled to fly around, and the concentration gradient measurement and maximum gradient direction determination process is repeated; and alternately executing the operation. Through the working mode that the two unmanned aerial vehicles alternately fly around and hover, rapid and efficient pollution source positioning is achieved.

The invention discloses a question and answer method based on medical institution internal knowledge base retrieval enhancement generation, electronic equipment and a storage medium. The method comprises the following steps: S100, constructing a medical institution internal knowledge base; s110, collecting an original document, preprocessing the original document, dividing the original document into text blocks, and storing the text blocks into a database; s120, establishing vectorization and indexing based on abstracts for the text blocks; s200, based on the questions of the user, performing retrieval to generate answers; s210, rewriting and expanding the original query to obtain an expanded query; s220, performing multi-path mixed retrieval, namely performing preliminary retrieval on the original query and the extended query to obtain N candidate text blocks; s230, rearrangement and careful selection are carried out, and M most relevant text blocks are selected by adopting a rearrangement model; and S240, answer generation and source tracing: integrating the original query, the M text blocks and the instruction large language model to generate a section of natural language question and answer to obtain an answer result. Accurate, efficient and traceable natural language question and answer service is provided for medical care and management personnel, and standardization, accuracy and efficiency of medical work are improved.

The invention belongs to the technical field of multi-source fusion, and particularly relates to a decision-making-level multi-source fusion gas leakage tracing method and a decision-making-level multi-source fusion gas leakage tracing device. The method comprises the following specific processes: data acquisition and space-time alignment: realizing decision-level multi-source acquisition on gas leakage data, and performing space-time alignment on the acquired data; preliminary leakage source positioning based on the evidence theory: processing the multi-source data by using the evidence theory, removing conflicting evidences, performing weak evidence weighting enhancement, and fusing the corrected evidences to obtain a region where leakage may occur; in the processing process, recording and obtaining a Jousseme distance; and accurate positioning of the leakage source: determining a dynamic radius based on combined adjustment of a detection error, time decay and a Jousseme distance, and performing weighted positioning by using a least square method to realize accurate source tracing of the leakage source.

The invention discloses an individualized precise brain regulation and control method and device based on dynamic brain network traceability, and relates to the technical field of electroencephalogram regulation and control. The method comprises the steps that resting state electroencephalogram signals of each lead of two resting state tasks are obtained; performing traceability analysis on the resting-state electroencephalogram signals of each lead of the two resting-state tasks, and calculating weighted phase lag indexes of lead pairs according to the traced resting-state data of the two leads in the pairwise lead pairs; constructing dynamic brain networks according to the weighted phase lag indexes of all lead pairs, performing modeling according to the first dynamic brain network and the second dynamic brain network, calculating an abnormal index of each brain region node, performing target stimulation based on the abnormal index of each brain region node by using a multi-channel stimulation controller, and performing target stimulation based on the abnormal index of each brain region node. According to the invention, the problems of inaccurate target spot positioning and single target spot in the existing cranial nerve regulation and control technology are solved.

The invention relates to the technical field of gantry crane dynamic operation safety testing, in particular to a crane dynamic operation safety testing method which comprises the following steps: step 1, multi-dimensional static calibration; step 2, fault coupling simulation; step 3, structured data processing; step 4, coupling judgment; step 5, source tracing; and step 6, carrying out targeted rectification. A multi-dimensional static calibration, fault coupling simulation and targeted rectification scheme is designed according to the structure and working conditions of cart one-way walking, main beam-supporting leg rigid welding, long track bearing, bilateral braking and open-air operation of the gantry crane. The limitation of insufficient adaptability to the integrated structure of the gantry crane in the prior art is broken through, and the problem of test distortion caused by neglecting equipment attributes is solved. Aiming at the characteristics that the box type girder of the gantry crane is large in torsional rigidity, small in torsional deformation and mostly caused by internal defects or uneven external stress, progressive traceability of internal defect detection, external load checking and connection state checking is adopted.

The invention discloses an illegal unmanned aerial vehicle source tracing method and system based on visible light, thermal imaging and radio frequency fingerprint fusion. The method comprises the following steps: synchronously acquiring target visible light, a thermal infrared image and a radio frequency signal; extracting appearance, thermal radiation and radio frequency fingerprint features in parallel; multi-modal features are dynamically weighted and fused through an adaptive cross-modal attention fusion network (ACMAF-Net), and unmanned aerial vehicle model identification and individual association are completed; reconstructing a three-dimensional track of the unmanned aerial vehicle by using multi-sensor observation data; the possible positions of a take-off point and an operator are deduced by combining track features and geographic information. The system comprises a multi-mode sensing subsystem, a data processing and fusion subsystem, a traceability analysis subsystem, a display control storage subsystem and the like. According to the invention, through three-mode complementary fusion, the limitation of a single technology is overcome, all-weather and high-precision detection identification and effective traceability of the illegal unmanned aerial vehicle are realized, and the low-altitude security and protection capability is remarkably improved.

The invention relates to the technical field of data processing, and discloses a medical knowledge base question answering method and system, electronic equipment and a storage medium. Performing mixed retrieval on the medical problem based on a medical knowledge base to obtain a set number of text blocks as retrieval results, constructing cue words based on the medical problem and the retrieval results, and generating a preliminary answer containing a preliminary source identifier through a large language model; source verification and source tracing processing is carried out on the preliminary answer to generate a final answer, and the source verification and source tracing processing comprises the steps of analyzing the preliminary answer to obtain a plurality of independent claim points, carrying out sentence-level and block-level correlation judgment on text blocks pointed by the claim points and the preliminary source identifier based on a reordering model, and executing source judgment based on a threshold value; generating a corresponding source mark; according to the medical knowledge retrieval method and system, the accuracy and integrity of medical knowledge retrieval can be improved, and the fact accuracy of answer generation is ensured.

The invention discloses an RAG question and answer traceability positioning method and device based on a text coordinate index, relates to the technical field of RAG and traceability positioning, and solves the problems that in the prior art, coordinate positioning precision is poor, and traceability in an original text is difficult after deconstruction. The method comprises the following steps of: analyzing and storing a text and corresponding coordinates into a mixed retrieval mode of Elasticsearch, meanwhile, carrying out vectorization storage on text fragments, deconstructing an original text paragraph or keyword in advance, storing the original text paragraph or keyword into mongoDB, carrying out vector similarity matching with Elasticsearch to find the original text when in use, and carrying out uniform format conversion on an input file, so that the extraction precision of the original text coordinates is higher, and the extraction efficiency is improved. The problem of tracing difficulty caused by RAG model text reconstruction is avoided to the greatest extent by utilizing an Elasticsearch technology, so that the most accurate tracing result can be obtained.

The application discloses a security vulnerability tracing analysis method and system for application software, and relates to the technical field of network security.The application solves the technical problem that it is difficult to quickly and accurately locate the source of the vulnerability and assess the risk level, thereby leading to the difficulty in formulating an effective security protection strategy.The application determines the source and propagation path of the vulnerability attack by performing correlation analysis on abnormal traffic, performs reverse DNS query and authenticity verification on the source IP address, and further improves the accuracy of the tracing result, thereby providing strong support for quickly locating the attack source.In terms of vulnerability risk level calculation, basic index scores, time index scores and environment index scores are calculated respectively, and the vulnerability score index is calculated in combination with respective weights, so that the actual situation is fully considered, the evaluation result is more in line with the actual risk, more accurate risk early warning, more targeted vulnerability repair and protection strategies can be provided, and the application software security protection level is improved.

This disclosure provides a method for unknown malicious source tracing analysis based on cross-platform behavior graph fusion. The method includes: inducing the collection of collaborative behaviors of multiple processes across platforms through an adversarial sandbox; semantically reorganizing the collected raw behavior data into a set of collaborative behavior events; constructing a subject-attribution-enhanced heterogeneous behavior graph by adding collaborative attribution edges; performing subject attribution determination on the subject-attribution-enhanced heterogeneous behavior graph based on a graph neural network model, merging the dispersed collaborative processes into a unified attack subject subgraph set; identifying high-risk behavior subgraphs and their stage propagation chains from the unified attack subject subgraph set; performing reverse tracing along the time sequence, resource transfer sequence, and control transfer sequence; and combining cross-platform same-source comparison to shrink into a single-source attack chain, outputting a source tracing conclusion. This method can solve the technical problems of multiple false starting points and the inability to form a single attack source attribution in the source tracing results.

The invention provides a large language model text traceability method based on virtual cue word embedding, and belongs to the technical field of artificial intelligence natural language processing and the technical field of digital content security, the method comprises the following steps: (1) constructing a chat input instruction text with a role format; (2) obtaining embedded vector input; (3) obtaining system cue word embedding representation; (4) obtaining a generated text with a user identity information watermark; and (5) recovering user identity information to realize traceability of the generated text. According to the method, on the premise that main body parameters of the large language model are not modified and model text generation quality and response efficiency are not affected, instant embedding of traceable identity information in the large language model generation process is achieved, meanwhile, robustness of embedded watermark information in a text transformation scene is guaranteed, and the method is suitable for large language model generation. Verifiable and reproducible identity identification information can be stably extracted from the generated text, and reliable technical support is provided for source tracing of the generated text.

This invention discloses a method and system for identifying abnormal behavior in wind farms based on multi-model cascaded calls, belonging to the field of abnormal behavior identification technology. The method includes: acquiring real-time video streams and determining scene semantic information; constructing an anomaly identification task graph; adaptively filtering and arranging a wind farm model library; obtaining a model cascaded call chain; performing multi-level progressive detection on the real-time video stream according to the model cascaded call chain; obtaining the output sequences of each model; performing error assessment and source tracing; and establishing a wind farm identification interference tree; performing multi-dimensional redundancy verification on the output sequences of each model; and establishing a wind farm identification redundancy verification space; and performing credibility enhancement processing on the output sequences of each model based on the wind farm identification interference tree and the wind farm identification redundancy verification space; and establishing a wind farm abnormal behavior heatmap. This invention solves the technical problem of insufficient accuracy in wind farm abnormal behavior identification in existing technologies, achieving the technical effect of improving the accuracy of wind farm abnormal behavior identification.

The invention relates to the technical field of aero-engines, and discloses a cartridge receiver quality problem tracing method, which improves the tracing analysis efficiency and reduces the influence of human subjective factors. A data confidence coefficient formula is adopted to quantify data reliability, so that interference of low-quality data on a traceability result is avoided, and data support is more scientific; through a Bayesian network posterior probability formula, quantitative sorting of problem reasons is realized, manual subjective judgment errors are reduced, and traceability precision is improved; by integrating data of a casing design stage, a processing stage and a detection stage, a'data island 'is broken, a constructed knowledge graph realizes historical experience reuse, and a systematic knowledge accumulation and reuse mechanism is established, so that a historical solution can be quickly called when similar problems appear repeatedly, and the tracing efficiency is improved; through dual verification of an evidence chain and a relation chain, accuracy and interpretability of problem positioning are provided.

This invention discloses a method and system for real-time tracing of live stream hotlinking based on multi-source heterogeneous signaling fusion. The method includes multi-source heterogeneous signaling acquisition, multi-source heterogeneous signaling feature extraction, construction of a hotlinking adaptation loss function, hotlinking tracing model design, parameter dynamic optimization, and real-time hotlinking determination of live streams. This invention belongs to the field of data processing, specifically referring to a method and system for real-time tracing of live stream hotlinking based on multi-source heterogeneous signaling fusion. This solution constructs a hotlinking adaptation loss function based on the arc-tangent power-law fallback characteristic to adapt to multi-signaling modal noise fluctuations; embeds a risk probability site inversion penalty term to effectively improve the reliability of real-time hotlinking tracing of live streams; and constructs a full-process prevention and control system: designs a parameter dynamic optimization strategy driven by inversion rate to dynamically adapt to data fluctuations, and establishes a structure-gradient dual-dimensional inversion control mechanism to avoid inversion risks throughout the entire link from model structure design to parameter gradient optimization, significantly improving the hotlinking tracing effect.

This invention provides a method and system for collaborative environmental monitoring and pollution source tracing using UAV swarms in complex terrain, relating to the fields of UAV environmental monitoring and pollution source tracing. The method includes: a ground control center planning layered and altitude-based tasks for the target area and sending them to each group containing master and sub-UAVs; the master UAV decomposing and encoding the sub-tasks and transmitting the results back; the ground-based optimization of flight paths being distributed; the sub-UAVs collecting pollution source data, which is then preprocessed by the master UAV and uploaded; ground-based analysis results being fed back; and the master UAV calculating the contribution value C using a source tracing model. en Once the source tracing is complete, the results are uploaded to the server, and the drone returns to base. The system includes a ground control center and at least one drone swarm; each swarm consists of a master drone and sub-drones. This invention can plan flight paths based on the geographical environment of the monitored target area, avoiding detection blind spots and conflicts, and coordinate actions based on the flight path, thus improving the convenience and accuracy of environmental monitoring and pollution source tracing.

This invention discloses a deep learning-based method for VOCs pollution feature analysis and source tracing. The method includes acquiring online monitoring time-series data of VOCs and synchronous meteorological parameters of the monitoring area, constructing a multidimensional input tensor after preprocessing, building a deep neural network model containing a spatiotemporal feature extraction module and a source apportionment regression module, extracting nonlinear chemical coupling features and temporal evolution features between VOCs components using a one-dimensional convolutional neural network and a bidirectional long short-term memory network, and weighting and fusing them using meteorological parameters, and finally decoding and outputting the source component spectrum and temporal contribution rate that conform to physical constraints through the source apportionment regression module to generate a pollution feature map. This invention effectively solves the problem that traditional receptor models struggle to handle nonlinear chemical transformations and complex meteorological couplings, improves the accuracy and physical interpretability of VOCs pollution source tracing, and has online update capabilities to adapt to environmental changes.

This application discloses a machine learning-based intelligent source tracing method and system for pesticide pollutants. The method includes: acquiring multi-source data and constructing a graph structure representing the migration and diffusion relationships of pollutants; calculating the deviation between the theoretical and actual migration times of pollutants between monitoring points based on historical monitoring data, using this deviation as a correction factor; constructing and training a machine learning model that integrates the correction factor for predicting the spatial distribution of pollutants in conventional scenarios; monitoring the data stream in real time, calculating data anomaly indicators using an anomaly detection model, determining an unknown pollution event when the indicator falls below a threshold, and automatically switching to physical model inversion and unknown substance structure analysis modes to deduce the location of the pollution source and identify the category of the unknown pollutant, respectively. This application enhances the response capability to unknown events such as sudden leaks through the synergistic collaboration of conventional source tracing and emergency unknown source tracing modes, and can be widely applied to intelligent early warning and source tracing of pesticide pollution in surface water.

The invention discloses a root tracing method and device for knowledge content, equipment and a readable storage medium. The method comprises the following steps: positioning at least one piece of difference information from target answer content uploaded by a target user; the root cause classification of each piece of difference information is determined, the root cause classification is the error type to which the information corresponding to the point losing point belongs, and the error type comprises information known and information unknown; and / or the error type comprises information missing and information redundancy; the root cause tracing result of the difference information under the error type is determined through the large language model, and the root cause tracing result of the difference information can be displayed to the user, so that the user can perform improvement based on the root cause tracing result of the difference information, and the analysis ability of the user is helped to be improved.

The invention provides a container mirror image source tracing and verification method based on a block chain, and the method achieves the transparent tracing of a whole process from the construction, updating to distribution of a mirror image through the non-tampering performance and distributed storage of the block chain. Generating a mirror image unique identifier through SHA-256, and performing uplink storage on metadata; and the IPFS is adopted for multi-copy redundant storage, so that the availability is improved, and a single-point fault is avoided. And the edge node verifies the integrity of the mirror image through on-chain hash, and carries out dynamic reward and punishment in combination with a reputation management model driven by an intelligent contract to inhibit malicious behaviors. The encryption and hybrid consensus algorithm is adopted to guarantee the communication security and performance, the mirror image credibility and the system robustness are remarkably improved, and the method is suitable for distributed scenes such as cloud computing and the Internet of Things.

The invention discloses an instant threat early warning and active traceability deterrence method and system based on a lightweight trigger mechanism in a power system network, and relates to the field of power system network security. According to the method, a light load trip line is preset at a key node, a suspicious attack log is generated in combination with a hidden Markov model, and multi-channel instant early warning is triggered; extracting multi-dimensional fingerprints of a browser, hardware and a special network weapon to construct a warehouse, and generating three types of flight fragments of deep acquisition, remote control and vulnerability repair for precise delivery; and performing confusion packing processing on the control software, and realizing active traceability deterrence in a matching / non-matching mode. According to the invention, threat identification accuracy and response speed are improved, an active defense closed loop is formed, and the method is suitable for security protection of a power system network.

The invention relates to a fire tracing and diffusion prediction method and system based on multi-source data fusion and a storage medium, and relates to the technical field of disaster monitoring and early warning. The fire tracing and diffusion prediction method comprises the following steps: collecting and cleaning multi-dimensional environment information of a disaster scene, and obtaining multi-modal environment data; constructing a disaster situation multi-dimensional map according to the multi-modal environment data, and dividing a field dangerous area in combination with a preset disaster situation threshold value; analyzing the disaster situation multi-dimensional map according to the real-time environment data collected by the intelligent carrier, determining the disaster situation category, and positioning the disaster situation generation source; monitoring environment change data of a disaster occurrence source, and constructing a disaster time sequence deduction model in combination with disaster categories; according to the disaster situation time sequence deduction model in combination with the disaster situation multi-dimensional map, pre-judging and determining the regional disaster situation level of the field dangerous region, and drawing a disaster situation diffusion rendering graph; and planning a rescue path of a disaster scene according to the position of the intelligent carrier in combination with the regional disaster level, and providing disaster early warning.

This application discloses a method and system for multimodal content source tracing analysis of social media propagation chains, including: collecting and preprocessing multimodal propagation data; encoding visual and semantic modal data through a deep feature extraction network and projecting it into a cross-modal alignment space to generate a fused semantic representation vector; calculating the weighted cosine distance through a modality adaptive attention mechanism to obtain the semantic fidelity attenuation, and combining it with platform encoding degradation estimation to obtain the net semantic attenuation; constructing a propagation dynamic graph based on propagation metadata with the net semantic attenuation as edge weights; aggregating propagation features through an edge-weight-aware graph convolutional network; tracing and locating the source node by performing reverse semantic gradient tracing through a source likelihood scoring network; and achieving specialized source tracing of adversarial mutation propagation through semantically invariant region detection and deep semantic structure correlation mining. This invention achieves deep coupling source tracing of content-level semantic information and propagation topology information.

The application provides a VOCs pollution source tracing method and device based on a convolutional neural network and a space-time-component coupled image and a storage medium. It relates to the technical field of atmospheric pollution. The method comprises a data set of sailing position coordinates, VOC species and concentrations; a multi-channel three-dimensional image tensor is constructed; a CNN model with global average pooling is used to extract feature information; an image data information is obtained by using a class activation map technology; the image data information is preprocessed to obtain activation intensity, which is used as a prior constraint to construct a constrained objective function model; the objective function is solved to obtain a source component spectrum matrix and a source contribution matrix. Through the application, the chemical fingerprint extracted by deep learning is used as a physical constraint, effectively solving the rotation ambiguity problem of the traditional PMF model solution, enabling dynamic inversion of the pollution source component spectrum changing over time, and realizing high-precision analysis and spatial positioning of the pollution source.

The invention discloses a traceability report generation method and device, electronic equipment and a storage medium, and the method comprises the steps: determining a target task for a research problem according to the research problem input by a user; creating a traceability knowledge base having an association relationship with the research problem; obtaining original text information of a to-be-executed step for executing the target task; metadata used for representing the logic relation between the original text information and the execution result of the to-be-executed step is determined; storing the original text information, the execution result and the metadata into a traceability knowledge base; and in response to a received traceability request of a user for the task report of the target task, extracting the original text information, the execution result and the metadata from the traceability knowledge base to generate a traceability report. According to the method, the original information, the intermediate execution result and the logic association metadata are synchronously stored in the task generation process, a complete mapping system from a fact manuscript to a final conclusion is constructed, and logic trace reserving and evidence path restoration of artificial intelligence generation content are realized.

The application discloses a clean room pollution source tracing method, system and equipment of adaptive grid dynamic reconstruction, comprising: acquiring the spatial characteristic value of the clean room, calculating the reference flow field data, and constructing the initial pollution sample library; the clean room space is divided into a primary grid and a secondary grid; real-time acquisition of the concentration time series data of the sensor, calculation of the concentration change gradient information, generation of a tertiary encryption grid; based on the tertiary encryption grid, multiplexing the reference flow field data, locally reconstructing the pollutant transport equation, and extracting the enhanced feature vector; inputting the mapping model, outputting the probability distribution and confidence of the pollution source position; after the tracing is completed, the tertiary encryption grid is released, the optimization parameters of the adaptive threshold are updated, the grid dynamic reconstruction is realized through the spatial distribution unevenness feature triggering mechanism, the problems of insufficient local precision or calculation redundancy of fixed grid division are avoided, the confidence evaluation is provided to support the risk decision, and the timeliness and accuracy of the clean room sudden pollution tracing are improved.

This invention provides a method for tracing the source of large language model text based on virtual prompt word embedding, belonging to the fields of artificial intelligence natural language processing technology and digital content security technology. The method includes: (1) constructing chat input command text with role format; (2) obtaining embedded vector input; (3) obtaining system prompt word embedding representation; (4) obtaining generated text with user identity information watermark; (5) restoring user identity information to achieve source tracing of generated text. This invention achieves real-time embedding of traceable identity information in the large language model generation process without modifying the main parameters of the large language model or affecting the model text generation quality and response efficiency. At the same time, it ensures the robustness of the embedded watermark information in text transformation scenarios, and can stably extract verifiable and reproducible identity information from the generated text, providing reliable technical support for tracing the source of generated text.