633 results about "Semantics" patented technology

The invention relates to the technical field of enterprise management, in particular to an intelligent decision support system and method based on cognitive logic and scenarized semanteme, and the method comprises the steps: obtaining enterprise decision data through multi-source data monitoring, carrying out the preprocessing, and generating cross-modal enterprise scene cognitive information; analyzing cross-modal association among the data, fusing a cognitive logical reasoning rule and a semantic understanding model, and constructing scenarized semantic decision fusion features; training an enterprise decision-making quality evaluation model based on historical cases, performing quality perception on a current decision-making scene, and outputting decision-making quality information; and an execution effect is judged in combination with an expected target, an optimization mechanism is started if the target is deviated, candidate schemes are generated by using a historical knowledge base and a multi-target optimization algorithm, an optimal solution is screened, and intelligent adjustment of a decision scheme is realized. According to the method, multi-source heterogeneous data and cognitive logic are fused, semantic understanding, dynamic evaluation and adaptive optimization capabilities of a decision system are enhanced, and scientificity and real-time performance of enterprise decision are improved.

The invention relates to the technical field of electrical digital data processing, and discloses a system conflict point intelligent review system based on a knowledge graph and a large language model, which comprises the following steps: constructing a dual-mode storage space containing an unstructured index and a structured logic graph, analyzing target text extraction features and triggering graph-based generation logic; converting the topological structure of the associated sub-atlas into a natural language instruction sequence to construct a forced logic constraint template, filling the template with a text, and inputting a pre-training language model to generate a verification result; according to the method, the discrete atlas topology is mapped into the linear logic constraint, random divergence of the generative model is restrained on the calculation principle, and precise decoupling and dynamic evolution of unstructured semantics and structured logic are achieved.

The invention discloses a knowledge question-answering method and system based on topic knowledge graph retrieval enhancement, and the method comprises the steps: firstly extracting a local topic represented in a triple form based on an original document through employing a large language model, carrying out the clustering, and generating a global topic triple set representing the global perspective of the whole document; secondly, on the basis of the global topic triple set, topic-guided entity and relation extraction is adopted, and a mixed knowledge graph is constructed; secondly, providing a semantic perception personalized PageRank algorithm, matching query semantics with semantics of edges in the mixed knowledge graph, and dynamically adjusting the weight of score propagation between nodes; and finally, designing a three-level progressive retrieval mechanism, retrieving multi-level information related to user query from the mixed knowledge graph, and inputting the multi-level information into the large language model to generate a final answer. According to the method, the semantic integrity and retrieval precision of the knowledge graph are remarkably improved, and the accuracy, comprehensiveness and enabling performance of generated answers are ensured.

The invention relates to a multi-modal virtual human real-time generation method based on an expression control embedding space, and belongs to the field of artificial intelligence. According to the method, an expression control embedding space is constructed and used for fusing voice semantics, a rhythm structure and multi-dimensional emotion information, and continuous and controllable multi-modal driving vectors are generated. The whole system has an end-to-end linkage mechanism from audio input to expression and action output. Semantic features, rhythm structures and emotional states jointly act on generation paths of lip and upper body postures and expression modalities, and all modal features are fused and expressed in a unified control space through a collaborative coding and time sequence alignment mechanism. And finally, a high-consistency and high-fidelity virtual human video is generated in real time through an output scheduling mechanism. The method has remarkable advantages in the aspects of modal fusion consistency, generation expression naturalness and emotion control flexibility, and can be widely applied to key scenes such as virtual human broadcasting, voice interaction agency and meta-universe digital identity construction.

The invention is suitable for the technical field of natural language processing, relates to a hierarchical memory and context awareness retrieval method of a large role model and a related product, and aims to solve the problems of limited model memory duration, insufficient retrieval correlation and insufficient personality consistency in a long dialogue. According to the invention, a short-term-middle-term-long-term three-level memory architecture is adopted, and a memory attenuation and migration mechanism is combined, so that dynamic metabolism of memory is realized; related memories are recalled accurately through a context semantics and role personality double-sensitive double-stage retrieval algorithm; relying on a personality-linked memory fusion and response generation strategy, the reply is ensured to fit personality setting; and a closed-loop adaptive learning mechanism of dialogue-memory-retrieval-generation-feedback is constructed, and the memory quality is continuously optimized. According to the method, the role large model can have the human-like continuous memory ability, the continuity, retrieval accuracy and personality consistency of long dialogues are remarkably improved, and the long-term personalized interaction requirements of scenes such as digital personality assistants and dialogue agents are met.

The invention belongs to the technical field of artificial intelligence task planning, and provides a task planning PDDL file automatic generation method based on natural language input, and the method comprises the steps: (10) constructing task environment description words: analyzing environment elements and participants, and designing a diversified scene framework; (20) building a knowledge enhancement fine tuning model: fusing a vector RAG knowledge base retrieval result and pre-training model parameters by the model, and inhibiting logic illusion; (30) PDDL file generation and knowledge constraint: driving the fine tuning model to output a correct PDDL file in combination with scene semantics and knowledge base rules; (40) multi-stage dynamic evaluation: executing a feasibility index verification instruction through a knowledge base rule matching degree; and (50) model closed-loop iterative optimization: updating RAG knowledge base content and model parameters based on execution feedback, and constructing a closed-loop iterative optimization model. The method has the beneficial effects that the instruction logic deviation rate is reduced through RAG knowledge base constraint, the complex scene instruction generation accuracy is improved, and the dynamic decision response time is shortened.

The invention provides a global and local feature fusion SMPL-X action-to-text generation method, and belongs to the field of artificial intelligence. The method comprises the following steps: preprocessing and coding an input SMPL-X action sequence into a double-flow action feature; through a cross-modal mapping module, the double-flow action features are mapped to a pre-trained large language model through independent projection branches, and global conditions and local action prefix embedding are obtained; through a text generation module, a decoder of a pre-trained large language model is used as a trunk network, text cue word embedding is extracted based on a text instruction given by a user, local action prefix embedding and text cue word embedding are spliced and then input into the decoder, and global conditions are injected into each layer of the decoder through a cross attention mechanism. And generating a description text in an autoregression mode. According to the method, the description text which is consistent with action semantics and has sufficient details can be stably and accurately generated, and the generation stability and the cross-scene applicability are improved when disturbance exists in the action sequence.

The invention discloses a pre-training language model output lexical element control method and electronic equipment, and relates to the field of computer vision and natural language processing, and the method comprises the steps: introducing a character region visual feature in an input image as a cross-modal constraint basis, positioning the character region of the input image, and extracting the visual feature, the generation process is strongly bound with input vision and character semantics, semantic anchor points are provided for follow-up temperature adjustment, and the semantic imbalance problem caused by indifference temperature control is avoided; the cross-modal consistency score of the output lexical elements and the character region is calculated, so that dynamic quantitative evaluation of the semantic suitability of each lexical element in the generation process is realized; and the sampling temperature is adaptively adjusted based on the cross-modal consistency score, so that the model can optimize the temperature setting in real time according to the semantic alignment condition in the generation process, and finally the dynamic and refined balance of the certainty and diversity / flexibility of the generation result is realized.

The invention relates to the technical field of voice semantics, can be applied to business scenes of financial science and technology, medical health and the like, and discloses a voice generation method, device and equipment based on pseudo-autoregression modeling and a medium, and the method comprises the steps: obtaining a training sample containing a text sequence, a prompt voice segment and a target semantic token sequence; performing continuous fragment mask training on the text-to-semantic model to obtain a pseudo-autoregression trained text-to-semantic model; generating candidate speech output by using the text-to-semantic model and the initial semantic-to-acoustic model which are subjected to pseudo-autoregression training, and constructing a preference data pair; updating the semantics-to-acoustics model based on the preference data pair to obtain a preference optimized semantics-to-acoustics model; and generating target voice output based on the target text and the target prompt voice. According to the method, the time sequence modeling capability of the model is enhanced through pseudo-autoregression training, and the voice generation quality is directly optimized through the preference data pair, so that the voice alignment precision and the subjective listening feeling performance are improved.

The invention relates to the technical field of passenger flow prediction, in particular to a passenger flow prediction method based on language model fusion, which is suitable for railway operation scenes such as train working diagram dynamic adjustment and passenger transport organization, and comprises the following steps: firstly, collecting passenger flow, holidays, weather and event multi-source data and preprocessing; in the first stage, passenger flow data is input by using time sequence models such as PatchTST and the like, and preliminary predicted values in a plurality of time periods in the future are output; and textualizing the predicted value, generating external semantic information, constructing a fusion Prompt input language large model, and optimizing the predicted value after comprehensive analysis. According to the passenger flow prediction method based on language model fusion, numerical trend and semantic collaborative modeling is realized, the prediction precision of holidays and holidays and emergency scenes is improved, the result interpretability is enhanced, and the prediction efficiency is improved. Generalization is high, the structure is flexible, and reliable support is provided for railway operation decision making.

The invention discloses a large language model robustness visual diagnosis method based on multi-dimensional features and adversarial attacks. The method aims to break through a mode that traditional evaluation only depends on a single aggregation index, and a multi-dimensional text feature exploration system covering vocabularies, syntax, semantics and a structural layer is constructed, and a large-scale antagonism disturbance mechanism and a task self-adaptive quantification strategy are combined. And generating structured feature-adversarial instruction-robustness diagnosis data comprising the cue word to be evaluated and the corpus. On the basis, an interactive visual analysis system is constructed, and through bidirectional linkage of a feature statistical view and a semantic projection view, a user is supported to realize progressive exploration from macroscopic feature screening to microscopic semantic attribution under the double view angles of cue words and corpora, so that a root cause causing the fragility of the model is deeply diagnosed. According to the method, the key feature combination influencing the stability of the model can be identified, so that a basis is provided for directional optimization of the model, and the diagnosis depth of robustness evaluation is improved.

The invention belongs to the technical field of education artificial intelligence and deep learning model optimization, and discloses a thinking tree enhancement-based test question knowledge system intelligent labeling method and system.The thinking tree enhancement-based test question knowledge system intelligent labeling method comprises the steps of collecting question data by introducing a thinking chain enhanced data generation mechanism, and generating a high-quality training sample; and a control mechanism is set, a question generation process is optimized, and the generated question is ensured to accord with teaching specifications in the aspects of knowledge point coverage, grade adaptability, difficulty matching and the like. Through a text encoder based on comparative learning optimization, the model can accurately carry out semantic alignment on questions and knowledge point labels, and label path information is fused through a double-coding mechanism, so that the hierarchical relationship between the labels and the capture capability of semantic dependence are improved. Finally, the constructed multi-label prediction model significantly improves the accuracy and generalization ability of knowledge point labeling, and can provide accurate support for personalized learning recommendation and teaching resource allocation.

The invention provides an automatic vulnerability detection rule generation method based on LLM and static code analysis, and relates to the technical field of network security and software engineering. The method comprises the following steps of: constructing unified structured representation of project code semantics and vulnerability features, classifying taint'source 'and'sink' based on a heuristic rule and LLM (Logical Language Model), and constructing a taint propagation path assisted by a large language model and performing risk assessment; automatically generating a rule based on a'generation-verification-repair 'closed loop mechanism; according to the method, uniform structured representation of code side information and vulnerability side information is constructed, accurate identification of specific project taint sources and sinks and effective paths in multiple environments is achieved, a path risk sorting mechanism is introduced to achieve priority detection of high-risk paths, an automatic rule of error feedback circulation aiming at correction is established, and the accuracy of detection is improved. And the correctness and the performability of the rule grammar are greatly improved, and logic vulnerabilities are reduced, so that the manual debugging and maintenance cost is reduced.

The invention relates to the technical field of data processing, in particular to a complex semantic-oriented business process multi-agent modeling and intelligent error correction method and system. The method comprises the following steps: firstly, based on a BPMN meta-model, disassembling business process elements, analyzing BPMN specifications, forming a structured framework of business process modeling, processing unstructured texts in parallel through an intelligent agent group, identifying business process key concepts, mapping the business process key concepts into semi-structured information, and generating a preliminary extraction result; verifying the integrity of the preliminary extraction result based on a BPMN specification, and generating an information integrity verification report; based on an information integrity verification report, missing information of the preliminary extraction result is retrieved through a knowledge base and complemented; performing secondary verification on the complemented extraction result, and correcting the complemented extraction result based on business logic and user requirements; and generating an XML code and a visual flow chart conforming to the BPMN specification based on the corrected extraction result. According to the method, the problems of semantic fracture and structure deficiency of the complex semantic text in business process modeling can be effectively solved.

The embodiment of the invention provides a detection report generation method and device, equipment and a medium, and relates to the technical field of artificial intelligence. The method comprises the steps of obtaining a to-be-detected text; inputting the to-be-detected text and the cue word corresponding to the to-be-detected text into the large language model to obtain a detection result output by the large language model; and generating a detection report according to the detection result. On one hand, the large language model has a generalization recognition capability, and can detect the deformation sensitive content which is subjected to processing such as alphabetic, harmonic replacement, traditional and simple conversion or insertion of special symbols and the like on the sensitive content, so that the missing report rate of the generated detection report is reduced; and on the other hand, the large language model has a semantic comprehension capability, can deeply understand context semantics and detect sensitive contents and generate a detection report according to the semantics, so that error interception in compliance scenes such as academic discussion and news reports is avoided, and the generation accuracy of the detection report is improved.

An assert statement generator employs a neural transformer model with attention to generate candidate assert statements for a unit test method that tests a focal method. The neural transformer model is pre-trained with source code programs and natural language text and fine-tuned with test-assert triplets. A test-assert triplet includes a source code snippet that includes: (1) a unit test method with an assert placeholder; (2) the focal method; and (3) a corresponding assert statement. In this manner, the neural transformer model is trained to learn the semantics and statistical properties of a natural language, the syntax of a programming language, and the relationships between the code elements of the programming language and the syntax of an assert statement.

The invention discloses an advertisement creativity matching method based on multi-modal content generation, and relates to the technical field of digital media content generation, and the method comprises the following steps: building a cross-modal time anchoring belt facing advertisement creativity matching, carrying out metaphor level decomposition on input text information, marking a symbol axis for image information, and carrying out data processing on the image information; obtaining an initial semantic boundary list; and constructing a culture fingerprint database according to the initial semantic boundary list, and mapping the territory taboo information and the brand symbol information into constraint tags to obtain a semantic guardrail set. According to the method, through cross-modal time anchoring and semantic boundary control, accurate correspondence of the text and the image in time and semantic levels is achieved, and it is ensured that generated content is clear in semantic meaning and adaptive in culture. In combination with breathing type phase traction and cultural fingerprint dynamic adjustment, multi-modal content rhythm and emotion are coordinated and unified, brand expression is kept stable, and the overall consistency and propagation effect of advertisement creativity are improved.

The invention belongs to the technical field of ancient book digital processing, and discloses a training knowledge-fused ancient book word sense disambiguation deep learning method, which comprises the following steps of: constructing a training knowledge graph; preprocessing ancient book image texts; constructing a deep semantic disambiguation model fusing training knowledge; word sense disambiguation reasoning and result output; and system integration and intelligent application interface design. The high-precision word sense disambiguation method oriented to ancient books and texts is constructed by fusing training knowledge and a deep learning technology, so that the recognition capability of complex semantic phenomena such as polysemy words, ancient and modern heterosemy and common and false characters is remarkably improved, the interpretability and field adaptability of the model are enhanced, the dependence on manual training is reduced, and the training efficiency is improved. The method realizes efficient understanding and intelligent processing of the semantics of the ancient books under the condition of low resources, and has good popularization and application prospects and culture inheritance value.

The invention relates to the technical field of computer data processing and network security, in particular to an illegal character automatic cleaning system, which comprises a rule compiling module for monitoring rule change, performing semantic fusion and topological mapping on a rule set, constructing a deterministic finite automaton and mapping the deterministic finite automaton into a state transition table; the state switching module is used for constructing a double-buffer context and operating lock-free switching through an atomic pointer to realize hot updating; the speculation execution module is used for carrying out vectorization pre-scanning based on a state transition table by utilizing single-instruction multi-data stream parallel loading, identifying a walk path, falling into a safe state for releasing and falling into a trap state for triggering external verification; the self-adaptive feedback module is used for counting trap state triggering frequency, generating a rule allergy report and dynamically adjusting the size of a read fragment; according to the method, the contradiction between rule flexibility and execution efficiency is solved, and high-performance cleaning based on speculative execution is realized.

The invention discloses a cross-modal conference information association retrieval method and system and a medium, and relates to the technical field of artificial intelligence, and the method comprises the following steps: carrying out feature extraction on obtained multi-source heterogeneous data to obtain multi-modal features, and uniformly mapping the multi-modal features to a first feature space of a preset dimension; in the first feature space, cross-modal deep fusion processing is performed on the multi-modal features, and a joint embedding space with consistent semantics is constructed according to the cross-modal deep fusion processing; constructing a vector index database based on a multi-modal feature vector in the joint embedding space, receiving a natural language query and mapping the natural language query to the joint embedding space, executing two-stage retrieval, and then obtaining a semantic fusion score based on calculated semantic fusion scores; multiplying a time sequence reward value based on the query time deviation and a dynamic reward value based on the core word matching degree to obtain a dynamic fusion score, and performing fusion sorting on the candidate set to obtain a final sorting result and an associated retrieval result; according to the method, refined sorting of the retrieval results is realized.

The invention relates to an audio recognition method and device based on marking and backtracking correction, and the method comprises the steps: segmenting a target audio, and enabling adjacent segments after segmentation to have a partial overlapping region; if a slice point exists in a non-mute segment of the target audio and the acoustic feature similarity of a preset number of frames before and after the slice point is smaller than a set similarity threshold value, the slice point is marked as a cut-off risk point, and the cut-off risk point is used for indicating that continuous semantics before and after the slice point has a cut-off risk; after a text corresponding to the audio of each slice is recognized through the speech recognition model, the language confidence degree of an overlapping area where the truncation risk point is located is determined, and the language confidence degree is used for indicating context language logic of the overlapping area; and if it is determined that the language confidence is lower than a set confidence threshold, correcting the recognition texts of the slices before and after the truncation risk point. According to the invention, the accuracy of audio recognition is 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 provides an intelligent quality inspection method and device based on RAG and a dual-enhancement mechanism, and the method comprises the steps: determining a retrieval text based on a transcriptional text of voice data to be subjected to quality inspection and an initial quality inspection result; based on the semantics of the retrieval text and the scene label of the voice data, retrieving from a preset knowledge base to obtain matched entries and the relevancy of the matched entries; whether an IDK enhancement mechanism is triggered or not is judged based on the matching items, if not, a direct preference optimization DPO model is called to conduct preference matching degree scoring on the initial quality inspection result, and the DPO matching degree of the initial quality inspection result is obtained; and determining the accuracy and confidence of the initial quality inspection result based on the correlation degree of the matched items, the confidence of the large model, the DPO matching degree and the accuracy of the historical similar cases, and determining a final quality inspection result of the voice data based on the accuracy and confidence. According to the method and the device provided by the invention, the result accuracy, the illusion resistance and the service adaptation efficiency of intelligent quality inspection are remarkably improved.

The invention discloses a content abstract generation method based on chapter structure analysis, and belongs to the technical field of natural language processing. The method comprises the steps that firstly, an original text is preprocessed, then text structure deep analysis is carried out, the text type of the text is recognized, an explicit / implicit text relation is extracted, and special symbols are introduced through a Prompt normal form for implicit text relation extraction to strengthen logic semantics; secondly, scoring sentences by adopting a double-path scoring mechanism in combination with a deep neural network model of chapter structure features and an optimized text sorting algorithm, and fusing scores through a logistic regression model; then, screening target sentences based on a chapter relation weighted secondary modulus function and a greedy algorithm, and finally, carrying out post-processing to generate an abstract. According to the abstract generation method, the chapter structure logic is deeply utilized, so that the problems of logic unsmoothness, information redundancy or key relation missing in the existing abstract generation are solved, and the semantic coherence and information integrity of the abstract are improved.

The invention discloses a physical feature perception large language model construction method for flow field understanding and generation, and belongs to the technical field of intelligent aerodynamics, multi-modal deep learning and computer vision crossing. High-level aerodynamic semantics and bottom-level physical feature learning of a flow field are decoupled through a double-codebook sharing mapping mechanism, and alignment of the high-level aerodynamic semantics and the bottom-level physical feature learning is kept at the same time. The discrete unified representation obtained through the token device is combined with a large language model, high-precision physical reconstruction and deep aerodynamic semantic understanding of the flow field image can be achieved at the same time under a single model framework, and the core problems that in the prior art, perception and generation tasks are split, and the fidelity of physical characteristics is low are effectively solved.

The application discloses a kind of methods for automatically analyzing and evaluating report, it is related to computer software and information processing technical field, including: based on graph neural network and visual-textual dual modal feature extractor, realize the adaptive analysis of complex non-standard structure report version, and the document is parsed into two-dimensional data table containing different classification dimensions;Based on BERT vectorization, etc. Adaptive clustering identifies the cluster structure of uneven semantic distribution, adopts the agglomerative hierarchical clustering to construct tree-shaped multi-granularity semantic hierarchy, realizes semantic aggregation;The application realizes the text content of non-standard form, the identification and processing of complex semantics;Overcome the defects of low efficiency, easy to make mistakes, subjective influence and only read specific format or specific location of text content, lack of flexibility and unable to handle complex semantics in prior art manual input.

The invention provides a knowledge graph and text fused railway equipment fault diagnosis method and system, and the method comprises the steps: inputting a natural language fault description, analyzing the natural language fault description into a text, preprocessing the text to obtain a text sequence, inputting the text sequence into a pre-training BERT, extracting a context semantic vector of each word, inputting the context semantic vector into a BiLSTM to capture a context dependency relationship and time sequence features, and carrying out the recognition of the text sequence. Generating a sequence feature vector of fused semantics; the method comprises the steps of constructing a knowledge graph, carrying out token-level embedding fusion on text semantic features and knowledge graph entities and relationships, carrying out sequence relationship modeling by utilizing BiLSTM, calculating weights of fused semantic features by utilizing an attention mechanism, obtaining weighted semantic vectors, executing Softmax classification, and outputting fault category labels and confidence coefficients thereof; and outputting a diagnosis result, reasons and recommended measures, and providing an explanation path of related knowledge nodes. According to the method, the interpretability of the fault diagnosis process and the reusability of knowledge are realized, and the diagnosis accuracy is improved.

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

The application discloses a robot cognitive development method based on ontology semantics, and comprises the following steps: constructing a robot article identification professional knowledge base based on attribute function and ontology information representation of article definition; information determination based on attribute discrimination and semantic search; and robot cognitive development based on attribute information addition. The application simulates the process of human memory, learning and cognition of articles based on an ontology semantic knowledge base, and through machine learning and sensor attribute information, the robot can actively cognize, learn, expand and accumulate learned knowledge and experience according to information data; the robot can continuously develop its cognitive ability through learning, automatically construct a robot article identification professional knowledge base of unknown articles, and based on a semantic structure of triplets, share knowledge and exchange operation logic between man and machine, so that the cognitive level of the robot is improved and the operation experience of an operator is improved.

The invention discloses a semantic and physical perception-based unmanned aerial vehicle navigation extension method during testing. The method comprises the following steps: S1, constructing multi-modal navigation input information of an unmanned aerial vehicle; s2, based on multi-modal navigation input information, judging whether a unique candidate navigation task target conforming to the description of the navigation task instruction exists in the current view image or not, and correspondingly constructing an initial candidate waypoint set; s3, generating corrected candidate waypoints and a corrected candidate waypoint set by using the visual language model; s4, performing multi-dimensional quantitative scoring on the corrected candidate waypoints in the corrected candidate waypoint set; and S5, selecting the corrected candidate waypoint with the highest multi-dimensional quantitative score as an optimal waypoint, and generating a track containing a continuous pose sequence based on the optimal waypoint to control the unmanned aerial vehicle. According to the method, the navigation planning and self-correction capability can be enhanced, and the navigation decision accuracy and reliability are effectively improved.