237 results about "Canonicalization" patented technology

The invention relates to the technical field of natural language processing, and discloses a knowledge base question and answer platform construction method based on a large language model, which comprises a knowledge acquisition module, a data preprocessing module, a text processing module, a vectorization module, a question understanding module, a mixed retrieval module, a prompt generation module, an answer generation module and an answer quality analysis module. A secondary inquiry processing module and a feedback learning module; according to the method, a semantic segmentation algorithm is combined with semantic retrieval and keyword retrieval, so that the flexibility is high; normalized prompts are constructed, input is performed according to correlation sorting, and the accuracy of answers is improved; multi-dimensional confidence evaluation is introduced, strict multi-layer security and compliance filtering is set, and the reliability of the system is ensured; the relevance of multiple rounds of dialogues is judged and complemented, so that interaction is more natural and efficient; knowledge is collected and updated in real time, a knowledge base and a retrieval strategy are continuously optimized, and a closed loop of data-application-feedback-tracing-optimization is formed.

The invention discloses a two-stage electric power vertical field large model fine tuning method based on supervised fine tuning and reinforcement learning, and relates to the field of artificial intelligence large model fine tuning training in the electric power field. Performing supervision and fine tuning on the language model according to the field knowledge guide prompt words; constructing a reinforcement learning objective function by using an importance sampling ratio and a cutting mechanism; and constructing an output reward model and a standardized dominant function, and performing fine tuning on the adjusted language model in the initial stage in combination with a reinforcement learning objective function to obtain a fine-tuned power vertical field language model. According to the method, an electric power professional data set is constructed, the advantages of supervision fine tuning and reinforcement learning are combined, a technical path of staged optimization is formed, a standardized output format of electric power problem reasoning is designed, and the interpretability and engineering practicability of a model reasoning result are ensured.

A system for real-time detection and mitigation of morphing malware in high-density edge networks, consisting of: a data acquisition unit configured to receive, normalize, and encode multimodal telemetry data streams originating from at least one of the following domains: network traffic, process behavior, system call sequences, binary instruction traces, and control flow graphs; the data acquisition unit is further configured to compute feature embeddings over sliding time windows and apply privacy-preserving redactions prior to storage; a generative neural processor that is operationally coupled to the data acquisition unit and configured to generate synthetic morphing malware variants by learning probabilistic transformations of previously observed malicious data representations, maintaining semantic functionality while varying structural and behavioral features; a discriminative neural processor trained adversarially with the generative neural processor, wherein the discriminative neural processor is configured to detect morphing malware by evaluating a probability distribution over multimodal telemetry embeddings and classifying anomalous process and flow behaviors in real time; a coordination processor that is communicatively connected to both the generative neural processor and the discriminative neural processor and is configured to orchestrate adversarial co-training, regulate detection thresholds, calculate reinforcement-based penalties for false negative results, and trigger countermeasures as soon as a detection confidence level exceeds a predefined adaptive threshold; a secure, system-integrated inference and enforcement unit configured to perform low-latency countermeasures at the network edge, including selective packet filtering, flow isolation, process interruption, or system microsegmentation, based on instructions from the coordinating processor; and a hardware-embedded security enclave that is embedded in the system and configured to store cryptographic keys, neural model parameters, and integrity affirmation data to ensure the confidentiality, authenticity, and immutability of model artifacts and policy configurations.

The invention relates to a geology vertical field large language model construction method and system fused with a knowledge graph, and the method comprises the steps: obtaining multi-source geology data, and carrying out the standardization processing, and obtaining a structured geological corpus; constructing a first knowledge graph and a vector index database supporting semantic retrieval based on the structured geological corpus; extracting first path information and constructing a first cue word, and performing fine tuning training on a pre-trained large language model according to the first cue word; and receiving an original text input by a user, and performing normalization processing to obtain first prompt information. And constructing a second cue word in the vector index database and the first knowledge graph based on the first cue information, and inputting the second cue word into a large language model subjected to fine tuning training to obtain a geoscience domain question and answer with a controlled structure. Compared with the prior art, the technical problem of semantic deviation and uncontrollability of question and answer generation content can be solved.

The invention relates to the technical field of OA examination and approval, in particular to a multi-modal file intelligent examination and approval method and system based on a large language model, and the method comprises the steps: configuring exclusive examination and approval templates for different examination and approval roles, and generating a role exclusive intelligent examination model; the method comprises the following steps: receiving a to-be-approved multi-modal file, analyzing file content, and outputting structured JSON (JavaScript Object Notation) data; inputting the structured JSON data and a configured approval template into a corresponding intelligent review model, performing step-by-step reasoning through a thinking chain prompt strategy, and outputting a structured review intermediate result; encoding the review intermediate result into a semantic vector, and retrieving Top-K similar historical approval cases from a knowledge base module based on a vector database to generate a reference suggestion set; fusing the review intermediate result with the reference suggestion set to generate a review result; based on the review result and the feedback of the approver, the standardized approval opinions are generated, and the accuracy and consistency of review are improved.

The invention belongs to the technical field of natural language processing and geographic information systems, and particularly relates to a region address standardization method based on knowledge graph enhanced retrieval. Cleaning and preprocessing the original address text input by the user; utilizing a fine-tuned large language model to identify a geographic entity and performing standardized expansion on variant expression to generate a query candidate set; entity linking and context retrieval are carried out based on the knowledge graph, and attributes, hierarchy and spatial topology information of associated entities are obtained; in combination with the original input and the map context, an enhanced retrieval query text is generated through large language model reconstruction; vectorized semantic retrieval is carried out through the fine-tuned embedding model, and a preliminary candidate address set is obtained; carrying out multi-dimensional refined sorting by adopting a resorting model; and generating a structured standard address by using a large language model, and outputting the structured standard address after multi-level verification. According to the method, the problems of ambiguity resolution, alias recognition and context understanding in address processing are solved, and the accuracy and robustness of address standardization are improved.

The invention provides a multi-modal data pairing method and system based on deep learning, and relates to the technical field of data processing, and the method comprises the steps: obtaining a video multi-frame sequence and a target text, and respectively extracting an overlapped frame group set and a standardized text sequence; performing spatio-temporal feature extraction and text dependency relationship coding to obtain a video time sequence vector sequence and a text vector sequence; executing cross-modal alignment search, and constructing a monotonic matching path set; calculating a semantic and action entity relationship consistency score of the paired elements on the path to obtain a comprehensive score; and determining an alignment relationship between the video and the text based on the optimal path. According to the method, accurate matching of the video and the text is realized, and the cross-modal retrieval efficiency is improved.

The invention relates to the technical field of chemical safety intelligent analysis, and discloses a chemical safety production safety intelligent analysis system based on a large language model, and the system comprises a semantic data construction module which is used for obtaining multi-source data and building a standardized database, a semantic anchor point library and a causal fragment set; the mechanism constraint modeling module is used for generating a mechanism constraint library and mapping the mechanism constraint library with the device unit; the semantic compiling constraint module is used for compiling the natural language intention into a formalized constraint set; the interlocking generation matching module is used for generating a semantic interlocking table and an action list; the check amplitude limiting decision module is used for determining a reserved action list, an amplitude limiting parameter and an execution sequence; the simulation verification and evaluation module is used for carrying out full-track simulation and outputting a feasibility result; and the deviation revision backfilling module is used for comparing the actual working condition with the simulation track and generating a revision proposal. According to the method, self-adaptive closed-loop optimization of semantic knowledge unified modeling and safety analysis in the chemical safety production process is realized.

The invention provides a file positioning management method and system based on artificial intelligence, and relates to the technical field of artificial intelligence. According to the method, files are collected from multiple sources and subjected to standardization processing, an element set is generated in combination with multi-modal analysis of texts, images, audios, videos and tables, cross-modal alignment is achieved through a semantic representation model, hierarchical indexes of semantics, keywords and relations are constructed, and a unique traceability identifier is generated; in the query stage, intention recognition and joint retrieval are carried out, a result subjected to permission verification and traceability information labeling is output, online optimization and incremental reconstruction are executed based on user feedback, and comprehensiveness, accuracy, traceability and self-adaptive optimization of file positioning are achieved.

The invention discloses a self-adaptive text extraction method and system based on artificial intelligence, and the method comprises the steps: carrying out the analysis of the document structure entropy of an example document set, quantifying the noise density, geometric distortion degree and background complexity of the example document set, and carrying out the self-adaptive selection of a preprocessing assembly line intensity grade according to the above; dynamically configuring image preprocessing parameters and AI recognition model parameters, and generating a recognition engine instance to output a preliminary recognition text; after regularized coarse screening extraction is carried out based on key field description, a multi-candidate generation strategy is started for low-confidence-coefficient candidate text fragments, a multi-person cooperative verification process is triggered for lower-confidence-coefficient fragments, finally all the fragments are processed through a text standardization module, and structured text extraction information is output. According to the method, accurate adaptation of processing intensity is achieved through document quality quantitative evaluation, the extraction accuracy and system robustness of complex heterogeneous documents are effectively improved through a multi-level confidence coefficient verification mechanism, and the identification error risk caused by image quality fluctuation or rule solidification is reduced.

The invention relates to the technical field of knowledge maps, in particular to a steel manufacturing knowledge map construction method based on multi-source heterogeneous data, which comprises the following steps: converting message data into a standardized semantic triple through a semantic mapping rule, and complementing missing attributes by using physical and chemical component balance logic to form a standardized entity; meanwhile, a semantic processor is used for extracting unstructured report text logic, cross-modal coupling is carried out on the unstructured report text logic and SEM image features, and a failure mechanism enhanced entity is generated. Then, full-process dynamic parameters are collected and regularized into equal-interval time sequence parameters; and finally, taking the normalized entity as an index feature and the mechanism enhanced entity as an association constraint, executing semantic fusion by utilizing a topological feature mapping engine, and constructing a dynamic weight graph structure by combining with a dynamic probability distribution correction operator iteration parameter weight. According to the method, efficient association and semantic organization of manufacturing whole-process multi-modal data are realized, and the retrieval precision and knowledge discovery capability of steel manufacturing data are greatly improved.

The invention discloses a life omics knowledge graph construction method and system based on a large language model, and relates to the technical field of computer data processing. The method comprises the following steps: acquiring and preprocessing multivariate life omics data, wherein the multivariate life omics data at least comprises an unstructured biomedical text; performing information extraction on the text data based on a large language model to obtain entity mention and relation description; standardizing and normalizing the entity mention and the relation description on the basis of a large language model in combination with an external knowledge base to obtain a standard knowledge triple; and storing the triple into a graph data storage system, and constructing the knowledge graph. According to the method, the powerful natural language understanding ability of the large language model is utilized, efficient information extraction is achieved through structured prompt or field fine tuning, the model is innovatively utilized for entity standardization of relation perception, and the accuracy of knowledge fusion is remarkably improved.

The invention relates to the technical field of informatization system development, in particular to a natural language-based complex data model generation method, which comprises the following steps of receiving business requirement input in a natural language form, preprocessing and normalizing input contents, processing ambiguity and incompleteness of requirements through a multi-round dialogue complementation mechanism, and generating a complex data model. Obtaining a complete and clear business demand description; and carrying out deep semantic analysis on the normalized business requirements by adopting a large language model. Aiming at the pain points that an existing data modeling technology is high in threshold, low in efficiency, difficult in quality guarantee, weak in integration adaptation and the like, the method has the remarkable advantage of multiple dimensions, non-technical background personnel can directly input service requirements through natural language interaction and deep semantic analysis on the premise of reducing the technical threshold, database knowledge and SQL specifications do not need to be elaborated, and the method is suitable for large-scale popularization and application. The cognitive gap of business and technology is spanned, the dependence on professional design talents is reduced, and the learning cycle of green hands is shortened.

The invention discloses an education question answering model routing method based on deep learning and a knowledge graph, and aims to solve the problems that the existing multi-model selection depends on text similarity, lacks knowledge points and first repair constraints, is easy to route by mistake and generates illusion. The method comprises the following steps: generating problem sub-graphs on an educational knowledge graph, carrying out relation perception graph neural network coding, constructing a model capability graph containing expression distribution and service attributes, forming masked cost under the constraints of types, privacy, time delay, cost and first repair coverage, and obtaining soft assignment by adopting unbalanced optimal transmission with entropy regularization; and interpretable routing is completed in combination with the meta-path attention subjected to interpretation consistency regularization training, so that the technical effects of reducing wrong routing and illusion rate, improving preferred correct routing rate and outputting meta-path interpretation are achieved.

The invention discloses an intelligent and automatic test case generation method based on a large language model, and relates to the technical field of large language model application, and the method comprises the following steps: collecting a software demand description file and an interface standardization file of a project, carrying out semantic analysis and mapping, generating a logic constraint set, and establishing a path mapping table; inputting the path mapping table into a large language model, generating a test scene and assertion, and executing semantic drift detection; when semantic drift is detected, a drift report is generated, a reverse correction process is executed, a patch prompt is generated, and the generation chain is executed again; and after the semantic distance of the prompt chain is detected to reach a convergence state, outputting a stable test case set. According to the method, the generation structure is controlled through the deterministic prompt sequence, so that the stability and controllability of test scene and assertion generation are realized; and through semantic drift detection and a reverse correction mechanism, real-time correction of semantic offset of the prompt chain is realized, and the accuracy and stability of the test case are improved.

The invention belongs to the technical field of computer data security, and particularly relates to a data contract-oriented strategy real-time compiling and consistency hot deployment method and system.The method comprises the steps that according to a strategy input by a user, strategy definition and standardized representation are based on formalized semantics, and a data contract-oriented strategy is compiled in real time; constructing a strategy abstract syntax tree with complete semantic information; inputting the abstract syntax tree into a multi-stage compilation optimization pipeline, and generating a high-performance intermediate representation (IR) code by using a cost model driven optimization decision mechanism; a policy distribution architecture based on a publishing-subscribing model is adopted, and an improved Gossip protocol and a Paxos atomic submission algorithm are combined to realize distributed consistency synchronization and real-time hot deployment of a policy; and efficient and safe execution of the strategy is realized through a register type micro virtual machine and a self-adaptive execution mode. The problems that in the prior art, execution performance has bottleneck, strategy updating is delayed, complex strategy supporting capacity is weak, and cluster strategy consistency is difficult to guarantee are solved.

The invention discloses a method and system for automatically generating headless test cases based on behavior description. The method comprises the following steps: obtaining an original behavior description text, and carrying out standardization processing on the original behavior description text to generate a standardized behavior description text; performing semantic analysis on the standardized behavior description text to generate a structured intermediate representation; matching each IR step in the structured intermediate representation with a preset step library, and generating a corresponding target step for each IR step; according to all the target steps, corresponding code templates are selected from a preset template library and combined in sequence, and headless automatic test cases are generated; and for an interface on which the headless automatic test case depends, selecting a matched Mock data version from the Mock snapshot warehouse according to the interface contract hash, and performing contract consistency verification. According to the method, through semantic analysis, step library matching and template combination, BDD description is fully automatically converted into executable scripts, and testers are liberated from repeated coding work.

The invention discloses a product recommendation method and device, computer equipment and a storage medium, belongs to the technical field of artificial intelligence, and is applied to a recommendation scene of financial products. According to the method, end-to-end data standardization, semantic analysis and privacy protection are carried out; performing disturbance and noise injection based on differential privacy; the method comprises the following steps: accurately capturing short-term fluctuation and long-term trend of customer demands by virtue of a multi-level time sequence algorithm, and extracting a typical demand change mode through segmented clustering and modeling; when significant demand fluctuation is detected, the online learning and deep reinforcement learning model can update a recommendation strategy in real time; the interpretability analysis module transparently presents a recommendation decision basis; user feedback is continuously absorbed through incremental learning, model parameters are continuously optimized, and finally it is guaranteed that an accurate personalized recommendation result is output under millisecond-level response. According to the method, the recommendation accuracy, the user satisfaction and the service conversion rate are remarkably improved, and the compliance and the expansibility are considered at the same time.

The invention relates to a test system based on a multi-modal knowledge base, belongs to the technical field of software testing, and solves the problems of insufficient reuse of historical project data and expert knowledge, lack of professional domain knowledge, illusion, low test case adoption rate and low practicability. The system comprises a knowledge base construction module used for constructing a hierarchical multi-mode professional field knowledge base; the preprocessing module is used for forming a standardized project development document and a software source code file of a to-be-tested project; the test case generation module is used for generating a test case and a software test description thereof based on the multi-modal professional domain knowledge base, the project development document of the project to be tested and the software source code file; and the test execution module prepares execution conditions according to the test case and the software test description thereof, and then executes the software test steps in sequence according to the input and operation description steps described by the test process to obtain an execution result of the test case. And knowledge-based and deep structured multiplexing of test assets is realized, and the test quality and efficiency are improved.

A system and method are disclosed for policy-constrained symbolic rendering of artificial intelligence outputs using deterministic evidence structures and authenticated display controls. Governed records derived from device, clinical, operational, environmental, or multi modal data are normalized, lineage-tracked, and assembled into deterministic evidence packs whose ranking, augmentation, pruning, and ordering are reproducible and traceable. Governed prompts constructed from these evidence packs are provided to internal or external artificial intelligence models under controlled, deterministic invocation parameters. Candidate model outputs—textual or multimodal—are symbolically interpreted, evaluated by a multi-policy engine, and constrained into compliant renderable forms. A cryptographic render-token architecture binds each constrained output to its governing evidence, policy state, model and tokenizer identity, viewport or session binding, environmental conditions, temporal validity, and decision lineage. A display gate renders only upon successful token validation and revocation checks. Correction-propagation mechanisms revoke affected tokens, perform deterministic replay, and issue superseding views for reliable, auditable use in safety-critical environments.

A method includes obtaining a plurality of training data sets each associated with a respective native language and includes a plurality of respective training data samples. For each respective training data sample of each training data set in the respective native language, the method includes transliterating the corresponding transcription in the respective native script into corresponding transliterated text representing the respective native language of the corresponding audio in a target script and associating the corresponding transliterated text in the target script with the corresponding audio in the respective native language to generate a respective normalized training data sample. The method also includes training, using the normalized training data samples, a multilingual end-to-end speech recognition model to predict speech recognition results in the target script for corresponding speech utterances spoken in any of the different native languages associated with the plurality of training data sets.

The invention relates to the technical field of code security, in particular to a data stream compliance detection method and device based on code semantic rule-driven LLM, and the method comprises the steps: carrying out the normalization processing of a to-be-detected source code: removing annotation, null lines, unifying branch circulation into an if-else structure, and establishing lt; code lines and line numbers gt; performing mapping; constructing a code semantic rule knowledge base; generating a code intermediate expression by using LLM; generating an execution path for the intermediate expression, constructing a symbol table, and reasoning and positioning non-compliant variables and corresponding code lines according to a data flow rule; generating a detection report; the device comprises a code standardization processing function module, a knowledge base construction function module and other function modules for achieving the method, manual intervention is not needed, C / C + + code compliance can be automatically detected, the detection precision and efficiency are remarkably improved, and non-compliance positions can be accurately traced.

The invention discloses a text duplicate checking method, a vehicle, a computer readable storage medium and a computer program product, and relates to the technical field of information processing. The method comprises the following steps: carrying out standardized preprocessing on an input text to be subjected to duplicate checking to obtain a current text fragment; performing semantic coding processing on the current text fragment by utilizing the target coding model to obtain a query vector; a similarity retrieval interface corresponding to the target search engine is called, a plurality of candidate vectors corresponding to the query vector are recalled from a semantic vector index, and the semantic vector index is constructed based on a historical text set and a graph index algorithm in the target search engine; sorting analysis is conducted on the multiple candidate vectors based on the multi-level duplicate judgment threshold values, a duplicate checking result is obtained, and the duplicate checking result is used for representing the duplicate situation between the historical texts corresponding to the multiple candidate vectors and the input text. The technical problems of low accuracy and low efficiency of a text duplicate checking method in related technologies are solved.

The invention belongs to the technical field of artificial intelligence and demand engineering crossing, and provides a demand standardization method and system based on a knowledge base and a large model and a storage medium, and the method comprises the steps: constructing a hierarchical knowledge base; preprocessing the original natural language demand text to generate a standardized demand fragment; dynamically retrieving a normal form evidence packet and a domain knowledge evidence packet associated with the standardized demand fragment from a hierarchical knowledge base; and inputting the constructed cue word into a large language model, generating an output result by the large language model, performing multi-demand consistency verification and integrity and traceability verification, and correcting and outputting a final result according to verification and verification results. According to the method, automatic demand defect detection, standard text generation and full-link tracing are realized, the automation level and compliance of high-safety system demand engineering are remarkably improved, and the method is suitable for the fields of aerospace, ships, rail transit and the like.

The invention relates to an assembly type bridge modeling method based on a large model and an RAG technology. The method comprises the steps that a bridge standard component library is established based on an assembly type bridge standard image set; calling a large language model, and generating standardized knowledge statements based on the bridge standard component library; a word embedding method is adopted to convert the normalized knowledge statement into vectors, and the vectors are stored in a vector database; converting an input natural statement query into a vector by adopting a word embedding method, and retrieving in a vector database to obtain a normalized knowledge statement related to the query; performing parameter analysis and optimization on the retrieved knowledge statements based on a large language model by utilizing an RAG technology so as to generate component parameters; and calling a component modeling algorithm based on each component parameter to generate a component three-dimensional model, and assembling the component three-dimensional model to obtain a fabricated bridge model. According to the method, component parameter sorting, knowledge statement intelligent combination, efficient semantic retrieval, parameter optimization, three-dimensional model automatic generation and the like are organically integrated, and the modeling efficiency and quality are improved.

The invention discloses a text classification method and device and a processor. A normalized text vector is obtained by preprocessing and coding an input text. Then utilizing a multi-layer structure in a pre-trained bad text recognition model, firstly carrying out first-level category classification on the input vector through a first Transform coding layer and a classification layer, and meanwhile, carrying out related semantic clustering through a first clustering layer to generate a first clustering text vector; then, a second Transform coding layer and a classification layer complete finer-grained secondary category classification based on the first clustering text vector, and continuous aggregation is performed through a second clustering layer to generate a second clustering text vector; and finally, the third Transform coding layer and the classification layer are combined with the second clustering vector to output a global classification result. And finally, all classification results are screened by setting a confidence threshold, high-confidence categories are reserved, and accurate classification results with clear levels are obtained.

The invention discloses a multi-platform data distribution method and system based on dynamic mapping, and relates to the technical field of computer data distribution.The method comprises the steps that a source manuscript is obtained, a unified skeleton structure is established, and the source manuscript is embedded into the unified skeleton structure; a difference patch is generated when a manuscript operation node is triggered, a standard difference set is formed through standardization and serialization processing, and a dependency graph is constructed based on the set. And selecting a minimum patch subset meeting the target platform data requirement in the dependency graph through a genetic algorithm, and applying the minimum patch subset to the unified skeleton structure to obtain an updated manuscript instance. And mapping the update instance into a target platform data mode field by field, splitting the update instance into a difference patch sub-sequence according to a dependency relationship, packaging the difference patch sub-sequence into batches in combination with quota constraint, and distributing the difference patch sub-sequence after adding an idempotent order number and a version number. After distribution, receipt information is collected, and batch strategy parameters are dynamically updated. Repeated adaptation logic is avoided, cross-platform consistency and compliance are ensured, and distribution efficiency and expandability are improved.

The invention discloses a logic knowledge construction method and device based on a large language model and logic programming, electronic equipment and a storage medium, and belongs to the field of artificial intelligence knowledge engineering. According to the method, automatic conversion from multi-source heterogeneous knowledge to machine reasonable logic knowledge is realized through combination of semantic understanding capability of a large language model and standardized processing of logic programming. The method specifically comprises the steps of obtaining unstructured / semi-structured original knowledge data, performing entity relationship recognition and predicate logic detection by utilizing a large language model to generate a preliminary logic expression, performing grammar normalization processing to form standard logic facts and rules, and constructing a consistent logic knowledge base through a conflict detection and resolution mechanism. According to the method, a mixed conversion framework of natural language and formalized logic is innovatively provided, and the maintainability and expandability of the knowledge base are remarkably improved by introducing timestamp metadata, a rule dependence graph and an incremental updating mechanism. The technology provides structured logic support for subsequent large model reasoning, and can be widely applied to intelligent questions and answers, decision support and other scenes requiring precise logic reasoning.

The invention belongs to the technical field of internet security, and particularly relates to a malicious task script detection method and system based on federal map learning, and the method comprises the steps: obtaining a PowerShell script, and carrying out the preprocessing of the PowerShell script, and obtaining a standardized abstract syntax tree; constructing a directed acyclic command flow graph based on the normalized abstract syntax tree, obtaining an adjacent matrix according to the command flow graph, and performing word embedding on the command flow graph to obtain a node feature matrix; performing GAT training on the node feature matrix and the adjacent matrix to obtain vector representation of the command flow graph, inputting the vector representation of the command flow graph into a preset federal learning detection model for detection, judging whether the script is a malicious script or not, outputting a judgment result, and deeply analyzing the script structure and semantics by constructing an abstract syntax tree and the command flow graph to obtain the malicious script. According to the method, wider confusion technologies and attack techniques including attacks without specific keywords can be detected, and the problem that attackers update the attack methods and cannot effectively detect the attacks is avoided.

The invention discloses a multi-source complex table-oriented trusted question and answer agent construction method and system, a question and answer agent is constructed based on a ReAct framework, at least a code interpreter tool and a document reading tool are used in the ReAct framework, and the multi-source complex table-oriented trusted question and answer agent construction method comprises the following steps: constructing a multi-mode complex table analysis tool; the analysis module is used for analyzing a non-editable unstructured table file into an editable structured table text; constructing a single-mode complex table cleaning tool for performing abstract modeling, identification extraction and structure reconstruction on a complex data structure of the structured table text to generate a normalized table which is simpler than an original table in structure; and constructing an illusion detection and repair tool for performing real-time illusion verification on the output content of the question and answer agent, and triggering a fallback regeneration mechanism when illusion is detected. According to the invention, through tool set expansion and multi-tool cooperation, an efficient and credible table question and answer agent is constructed.