47 results about "Cognitive model" patented technology

The invention relates to the technical field of education, and discloses an online learning environment-oriented education interaction behavior recognition system, which is characterized in that a standardized cognitive model is constructed offline based on interaction data of a high-performance learning group; in the online stage, interaction events of individual learners are obtained in real time, high-dimensional cognitive trajectories representing cognitive states of the learners are generated, a trajectory dynamics analysis module executes double reference analysis, and internal dynamics indexes representing stability of the trajectories are obtained by calculating the maximum Lyapunov index, recognizing phase change points and the like; the external deviation measure of the difference between the trajectory and the normalized cognitive model is obtained by calculating the state deviation degree, the path efficiency and the like, and finally, the cognitive mode diagnosis module integrates the indexes and generates a structured diagnosis description according to a rule base. According to the method, through combination of internal dynamics and an external reference system, deep quantification and fine identification of the cognitive process of the learner can be realized, and an objective basis is provided for adaptive intervention.

The invention relates to the technical field of semantic cognition modeling and natural language interaction, in particular to an electric power digital twin question-answer interaction generation method based on semantic cognition modeling, which comprises the following steps of: accessing a power grid topological data flow in real time, analyzing equipment hierarchy, power supply attribution and physical connection relationship, and constructing a dynamic topological knowledge graph; the method comprises the following steps of: establishing a topological change monitoring channel, mapping the topological change monitoring channel to a semantic cognition model, embedding a topological hierarchy field and an association constraint rule, generating an enhanced semantic cognition model with topological association constraints, and establishing a topological change monitoring channel to update an association relationship in real time; receiving user natural language query, identifying keywords, extracting power supply attribution constraint rules, filtering invalid entities, and converting implicit topological semantics into combined query statements; and executing query to obtain topological associated data, verifying the power supply attribution consistency of a result set, filtering homonymy matching errors, triggering model updating and secondary verification during topology adjustment, and outputting question and answer responses conforming to topology constraints, so that the interaction accuracy and timeliness are improved.

The invention is notably directed to a computer-implemented method of clustering anomalies detected in a computerized system. The proposed method makes use of an unsupervised cognitive model, executed based on input datasets to obtain clusters of anomalies. The method accesses input datasets, which correspond to detected anomalies of the computerized system. These anomalies span respective time windows. Each input dataset comprises a set of timeseries of key performance indicators. The key performance indicators of each input dataset extend over a respective time window. That is, each anomaly corresponds to a respective time window. This model includes a first stage, which includes an encoder designed to learn fixed-size representations of input datasets, and a second stage, which is a clustering stage. The model is executed based on the input datasets accessed, the first stage learning fixed-size representations of the input datasets and the second stage clustering the learned representations.

The invention discloses a submission message generation method and system. The method comprises the following steps: firstly, constructing a multi-role developer cognitive map, and simulating information attention key points and cognitive modes of different roles; establishing cognitive models and decision preferences of different role developers based on the acknowledgement science theory and the large-scale development behavior data; then designing a code change importance perception algorithm based on an attention mechanism, simulating a human visual scanning and importance judgment process, and dynamically identifying key change contents concerned by a developer; and finally, establishing a context-aware message personalized generation framework, integrating multi-dimensional context information such as project history, team culture and personal preference, and adaptively adjusting a generation strategy and an output format of a submitted message. According to the method, the CMG problem is redefined from a'text generation task 'to a'cognitive assistance system', and an intelligent submission message really conforming to human cognitive habits is generated through a cognitive process and a decision mechanism of a deep modeling developer.

The invention provides a multi-personality decision simulation and game method for military artificial intelligence. Simulation of an AI opponent model on complex human behaviors is enhanced through a DIKWP semantic framework. The system firstly configures personality parameters based on a psychological model (such as MBTI, big five personalities), and maps the personality parameters to an AI cognitive model to construct a behavior preference model; then, a noise injection algorithm is adopted to simulate emotion and cognitive deviation, and unique deviation of different personalities in decision making is reproduced; the multi-agent self-gaming engine enables AIs with different personality characteristics to repeatedly game in a red-blue confrontation environment, optimizes a strategy and evaluates winning rate distribution. And the decision fusion analysis module summarizes the data, generates personality performance comparison and behavior pedigree diagrams under each tactical situation, and predicts opponent behaviors. The DIKWP semantic integration interface ensures that a personality simulation result can be seamlessly connected with a knowledge graph and a strategic intention module of an existing AI system. According to the method, AI opponent modeling is enriched on the semantic level, the decision simulation degree of a real commander is improved, and as a supplement of an existing AI simulation system, the ability of military AI in the aspects of strategic pre-judgment and cognitive game is enhanced.

The invention discloses a semantic planning and behavior generation system driven by autonomous consciousness. According to the system, a DIKWP mesh cognitive model is adopted, an intention / purpose semantic layer is embedded into an AI internal architecture, and autonomous decision making is achieved through five-layer semantic representation of data, information, knowledge, wisdom and intention. The system comprises a perception module, an intention generation module, a semantic planning module, a meta-cognition monitoring module, an execution module and the like. The AI can set a task target by itself and dynamically plan an action scheme based on a semantic state. Under the action of a double-circulation element cognitive mechanism, the AI can monitor the deviation between a behavior and a target in real time and automatically correct a plan. Compared with a traditional black box model, each step of decision of the AI has a traceable semantic basis and interpretation which can be understood by human beings, and the interpretability and controllability of the system are improved. The autonomous planning system can automatically adjust strategies according to environmental changes, ensures that AI behaviors always accord with expected targets and ethical safety constraints, is widely applicable to the fields of service robots, automatic driving, intelligent assistants and the like, and has important application values and commercial prospects.

The invention provides a method and a system for constructing a personal digital awareness mirror image and realizing cross-generation intelligent inheritance, and solves the problems of cross-generation knowledge inheritance and the like, and the method comprises the following steps: S1, multi-modal data acquisition and double-track recording; s2, constructing a personalized digital behavior cognitive model; s3, model packaging and conditional access are carried out; and S4, multifunctional mode activation and interaction. The method has the advantages of cross-generation knowledge inheritance, good digital community treatment effect and the like.

The invention relates to the technical field of video processing, in particular to an outdoor sports teaching method, device and equipment and a storage medium, and the method comprises the steps: inputting initial multi-modal teaching data into a reinforcement learning task model for processing, and obtaining a target teaching video; comparing the target teaching video with the student action video to obtain a comparison result, and generating a teaching suggestion based on the comparison result; the training of the reinforcement learning task model comprises the following steps: training a basic learning cognitive model based on initial multi-modal teaching data to obtain a multi-modal closed-loop simulation generation model; processing the initial multi-modal teaching data based on a multi-modal closed-loop simulation generation model to obtain a to-be-compared teaching video; comparing the to-be-compared teaching video with the model teaching video to determine a target point location action video segment in the model teaching video; and training a multi-modal closed-loop simulation generation model based on the target point location action video segment and the initial multi-modal teaching data to obtain a reinforcement learning task model. And the accuracy of video output and teaching suggestions of the model can be improved.

The application discloses a story interactive active questioning method and system based on cognitive characteristics of children, and relates to the technical field of artificial intelligence. The method comprises the following steps: obtaining story content; performing multi-modal analysis on the story content to obtain an analysis result; constructing a four-dimensional cognitive model of a child according to the analysis result and historical interaction data of the child; selecting a target problem level in preset multi-level problems according to the four-dimensional cognitive model; generating an adaptive question corresponding to the target problem level; and updating the four-dimensional cognitive model and the multi-level problems according to the answer and emotional feedback of the child to the adaptive question. The application first analyzes the story content and then generates a corresponding adaptive question in combination with the cognition of the child, so that the generated story question matches the cognition of the child, the problems beyond the cognition of the child are reduced, the adaptability is improved, and the learning and understanding effect of the child is improved.

The invention discloses a behavior tree generation method and device, electronic equipment, a storage medium and a program product, and relates to the technical field of computers.The method comprises the steps that an initial cue word corresponding to a virtual object and a target game environment where the virtual object is located are obtained; inputting the initial cue word into a pre-trained decision model to generate a first behavior tree corresponding to the virtual object; deploying the first behavior tree to a target game environment to obtain game match data; inputting the game playing data into a pre-trained cognitive model, and generating data description information corresponding to the game playing data; and optimizing the first behavior tree based on the data description information to obtain a target behavior tree of the virtual object in the target game environment. By implementing the technical scheme of the invention, automatic generation and optimization of the game virtual object behavior tree can be realized based on collaborative closed-loop iteration of the decision model and the cognitive model, and the intelligent level and the generation efficiency of the game virtual object are improved.

The invention discloses a voice image facial expression acquisition method based on a visual large model, and relates to the field of computer vision and intelligent cabin man-machine interaction. Uploading the collected video stream to a cloud server deployed with an artificial intelligence cognitive model based on a visual large model; receiving structured member information which is fed back by the cloud and contains member genders, ages and expression probabilities; executing multi-member optimization logic based on the information, and locking a target interaction object and extracting facial features by calculating a sorting score; and performing deformation on the parameterized model by utilizing facial features to construct a personalized grid, and converting real-time expression data into a weight to drive the grid to displace so as to generate a synchronous voice image. According to the method, accurate locking and high-fidelity expression driving of interaction objects in a complex scene are realized through an end-cloud collaboration and optimization strategy.

The invention discloses a visual shared whiteboard system and method for solving a group collaboration mathematical problem, and belongs to the technical field of computer science and human-computer interaction crossing. The system comprises a multi-modal fusion input module used for converting input such as handwriting and voice into a structured mathematical object; the real-time semantic understanding and knowledge graph engine is used for constructing a collaborative cognitive model; the collaborative process management module based on the directed hypergraph is used for carrying out formalized modeling and navigation on a nonlinear problem solving process; the context-aware intelligent auxiliary module is used for providing verification, recommendation and guiding consensus achievement; and a multi-user cooperation engine ensures stable cooperation experience. According to the method, the problems of professional expression, real-time cooperation splitting, group thinking process management difficulty and the like are solved, and a unified, intelligent and efficient integrated cooperation environment is provided for group solving of complex mathematical problems.

The invention relates to a multi-stage cognitive modeling and parameter optimization method based on a visual pairing comparison task, and aims to solve the technical problems of incomplete cognitive process modeling, weak crowd distinguishing ability, task parameter empirical and the like in the existing VPC task evaluation technology. Precise evaluation of the cognitive function and optimization design of task parameters are achieved through a five-step method, wherein VPC task parameterization design and eye movement data collection are carried out; constructing a'familiarity accumulation-memory attenuation-novelty attention and utility 'three-stage cognitive model; performing backstepping on individual free parameters based on maximum likelihood estimation; carrying out multi-population cognitive parameter distribution modeling and subtype division; and task parameter closed-loop optimization based on simulation and efficiency discrimination. According to the method, quantitative distinguishing of learning efficiency, memory stability and novel preferences is achieved, cognitive parameter recovery precision and multi-crowd distinguishing efficiency are improved, a standardized and transplantable technical path is provided for early screening of cognitive impairment and crowd typing, and the method is suitable for cognitive neuroscience research and clinical cognitive evaluation scenes.

This invention provides an information processing device that enables the modeling of individual behavior from both external and internal perspectives. [Solution] The information processing device is an information processing device that generates a judgment model of the behavior of traffic participants when crossing a traffic environment and executes a simulation, and comprises: an acquisition unit that acquires characteristic information that characterizes traffic participants, which is the premise of the judgment model; a first generation unit that, when simulating the behavior of traffic participants when crossing a traffic environment, generates a behavior model of traffic participants as a reaction to the understanding of the surrounding situation as an external factor when crossing a traffic environment, based on the characteristic information; a second generation unit that, when simulating the behavior of traffic participants when crossing a traffic environment, generates a cognitive model of traffic participants as a reaction to the understanding of the surrounding situation as an internal factor when crossing a traffic environment, based on the characteristic information; and a third generation unit that generates a judgment model of the behavioral judgment of traffic participants when crossing a traffic environment, based on the behavior model and the cognitive model.

This disclosure provides a cognitive model training and exercise recommendation method, apparatus, device, and medium, relating to the field of data processing technology, and particularly to the fields of artificial intelligence and deep learning technology. The specific implementation scheme is as follows: Cognitive Model Training Method: From the historical answer interaction records of each sample exercise, multiple interaction representation vectors between each sample exercise and the tester are extracted. The training set and validation set of each sample exercise respectively include partial interaction representation vectors corresponding to each sample exercise. Based on the training set of each sample exercise, an inner layer training is performed on a preset cognitive model. Based on the validation set of multiple sample exercises, an outer layer training is performed on the preset cognitive model after the inner layer training. Exercise Recommendation Method: The representation parameters of each sample exercise obtained after training the preset cognitive model according to the above method are used to initialize the representation parameters of each sample exercise in the exercise library. Based on the representation parameters of each exercise in the exercise library, exercises are recommended to the target user.

The embodiment of the invention discloses a multi-semantic theme fused smart education cognitive terminal control method and device. A specific embodiment of the method comprises the following steps: preprocessing an original education dialogue text set to obtain a processed text set and an evaluated text set; performing topic clustering on the processed text set to obtain a sub-topic text set; performing multi-topic fusion on the sub-topic text set to obtain a topic text set; updating the initial education cognition model to obtain an education cognition model; generating a real-time theme identification set based on the education cognition model and the real-time education conversation; based on the evaluation text set, determining theme indexes of the real-time theme identification set; and according to the real-time theme identification set, generating an education control instruction, and according to the education control instruction, controlling the intelligent education cognition terminal to adjust the output content and the teaching aid transmission device to perform teaching aid transmission. According to the embodiment, response time can be reduced, and wrong transmission of teaching aids is avoided.

The application discloses a kind of intelligent prosthesis control method and system with feedback mechanism, and the application relates to intelligent prosthesis control technical field, comprising the following steps: S1: acquisition step: through the sensor cluster of prosthesis foot bottom, ankle joint and knee joint deployment, real-time acquisition terrain feature, joint motion parameter and residual limb electromyogram signal.The intelligent prosthesis control method and system with feedback mechanism, through the double driving architecture of biomechanics cognitive model and reinforcement learning strategy model, in combination with the double verification mechanism of mutual inspection arbitration unit, realizes the adaptive accurate matching of feedback signal type, intensity and timing under complex dynamic environment.Biomechanics model ensures that feedback parameters meet clinical safety boundary, reinforcement learning model optimizes environmental adaptability in real time, mutual inspection arbitration resolves model conflict through physiological tolerance screening and control stability prediction, solves the problems of control misalignment, too high user cognitive load and walking confidence decline caused by traditional static rule feedback.

The invention discloses a human-based interaction system for artificial awareness cognition education. The system comprises an awareness cognition guiding engine, a human teacher cooperative training module, an interpretable behavior demonstration system, a cognition deviation reminding mechanism and a human-based intention driving feedback module. Wherein the guide engine extracts semantic expression, reasoning process and destination driving information of an artificial intelligence system based on a DIKWP multilayer cognitive model, and calculates information integration quantity phi of each cognitive module in combination with an information integration theory (IIT) so as to identify and mark a consciousness response stage with high-integration information and generate visual cognitive path interpretation; the teacher cooperative training module provides an interactive task release and hierarchical result feedback labeling mechanism to realize man-machine cooperative training; the behavior demonstration system synchronously displays cognitive state evolution in a graphical interface through a concept space orbit and a semantic space orbit; the deviation reminding mechanism monitors the logic consistency and the target matching degree and provides deviation correction interaction; the intention-driven feedback module collects the target intention, the interaction rhythm and the emotional response of the user to dynamically adjust the presentation mode and the interaction strategy, so that the interpretability, the trainability and the controllability of the artificial intelligence cognitive process are improved.

This application provides a control method and computing device for an embodied intelligent agent, relating to the field of artificial intelligence technology, to improve the accuracy and success rate of an embodied intelligent agent when performing a task. The method includes: inputting the task to be performed and initial context-aware information of the embodied intelligent agent into a readable cognitive model; determining the causal rules corresponding to the task to be performed through the readable cognitive model; the readable cognitive model is used to determine the causal rules that must be followed when performing the task in the current environment based on causal knowledge learned from multiple historical interaction data; inputting the task to be performed, initial context-aware information, and causal rules into an instruction generation model; determining a target action sequence through the instruction generation model; the instruction generation model is used to determine the action sequence that can complete the task in the current environment in a manner consistent with the causal rules; and controlling the embodied intelligent agent to perform actions according to the action order indicated in the target action sequence.

The invention relates to the technical field of tourism recommendation platforms, and discloses a knowledge graph-based tourism recommendation platform, which comprises a multi-role cognitive portrait construction module for constructing a structured multi-role cognitive portrait; the cognitive enhanced dynamic knowledge graph construction module is used for marking specific requirements of tourism resource entities on cognitive competence of the user; the multi-role demand fusion and conflict detection module is used for identifying potential conflicts and generating a coordination strategy; the cognitive adaptability itinerary generation and optimization module is used for generating an itinerary scheme giving consideration to cognitive demands of all members; and the experience feedback and cognitive model iteration module is used for calibrating cognitive portraits and resource cognitive attributes. According to the scheme, the dynamic knowledge graph fusing the user cognitive portraits and the resource cognitive attributes is constructed, refined modeling of multi-role group travel demands is achieved, the elastic travel taking consideration of cognitive suitability, space-time constraint and budget is generated, the model and resource annotation are continuously optimized by means of feedback closed loop, and the user experience is improved. And the heterogeneous group participation degree and the stroke feasibility are improved.

The invention provides a model construction method and system for a communication interference cognitive system. The method comprises the following steps: constructing a communication interference cognitive reference set containing interference scene association features and interference response reference features; then, interference feature correlation modeling is carried out based on the communication interference cognitive reference set to form a correlation map; implementing cognitive model structure adaptation according to the correlation map to generate an initial interference cognitive model; carrying out interference scene adaptive iterative optimization on the initial model to form an optimized target interference cognitive model; and finally, outputting the target model for interference identification and response processing of the communication interference cognitive system. According to the method, the interference characteristics are comprehensively captured, the incidence relation is deeply mined, the model structure is continuously optimized, and the accuracy and adaptability of a communication interference cognitive system in a complex interference environment are effectively improved.

The invention discloses an intelligent paper marking method and device for open education drawing questions, equipment and a storage medium, and the method comprises the steps: inputting a collected target answer sheet image into a preset detection model, and determining bounding box coordinates and question type labels corresponding to the open education drawing questions included in the target answer sheet image; according to the bounding box coordinates, cutting in the target answer sheet image to obtain a target answer area sub-image marked with a question type label; a chain thinking reasoning prompt template corresponding to the question type label is dynamically loaded, a multi-mode-based human-like cognition model is called, reasoning is executed in a target mode, and an implicit reasoning chain is generated; and based on an implicit reasoning chain, verifying the target answer area sub-graph to obtain a scoring result. Therefore, through mutual cooperation of the preset detection model and the multi-modal human-like cognitive model, intelligent paper marking of open drawing questions is realized, and the fairness and efficiency of evaluation are improved.

The invention discloses a cross-culture multi-industry ability semantic migration engine, and relates to application of artificial intelligence in the field of human resources. The engine constructs a global ability semantic graph and a neutral semantic presentation layer based on a DIKWP cognitive model and a knowledge graph technology, and performs equivalent reasoning and matching on skills of applicants among different language cultures and industry ontologies. The system can identify the essence corresponding relation of the capabilities behind different expressions, and cross-industry skill migration and cross-culture semantic alignment are realized. Accurate expression of each culture semantic is ensured through a semantic sovereignty principle, and proper abstraction is introduced by applying a consciousness' BUG 'theory to improve matching efficiency.

This invention discloses a system and method belonging to the field of smart home technology. The method includes the following steps: constructing a dynamic influence relationship hypergraph; triggering model verification logic; generating a virtual probe task based on the triggering result of the model verification logic and obtaining the expected influence propagation path; executing the virtual probe task while simultaneously collecting actual response data; comparing and analyzing the collected actual response data with the expected state in the dynamic influence relationship hypergraph, and updating and correcting accordingly; generating the final joint control command and sending it to the smart home execution device. This invention, by constructing a device relationship cognitive model that can dynamically perceive and self-correct, achieves the effect of ensuring the long-term accuracy and reliability of smart home automation joint control, solving the problem of control failure caused by the inability of static knowledge models in existing technologies to adapt to the dynamic evolution of the real home environment.

The invention provides an artificial intelligence model cognitive evaluation question generation method and system based on a DIKWP five-layer cognitive model (data, information, knowledge, wisdom and intention). By constructing a cognitive layer and AI capability dimension mapping table and a DIKWP * DIKWP semantic interaction matrix, a multi-language multi-mode engine is utilized to automatically generate test questions covering cognition of each layer, and question stems, options and reasoning prompts are output in a template form. The system adopts a dynamic sampling algorithm to regulate and control difficulty distribution, subjects are attached with labels such as cognitive categories, ability dimensions, difficulty, standard answers and the like, and evaluation and training reuse is supported. And a built-in verification mechanism ensures consistent logic, clear expression and reliable evaluation. Experiments prove that the method can quickly construct a high-quality sample library, realizes full-link quantitative evaluation of a large model from perception to intention, and improves the standardization and diagnostic value of white-box evaluation.

The invention discloses a traffic incident monitoring and handling method and device, equipment and a storage medium. Comprising the following steps: constructing a scene state vector based on system multi-source data, and inputting the scene state vector into a preset monitoring cognitive model to obtain risk values of various types of events and key scene elements of the events; when the risk value of the type of event is greater than a preset threshold value, performing high-risk event early warning, and outputting key scene element prompt information of the high-risk type of event; generating a feasible disposal strategy candidate set of the high-risk events based on system constraints, and performing optimization solution by adopting a preset disposal cognitive model to obtain an optimal disposal strategy; and decomposing the optimal processing strategy into specific control instructions, and issuing the specific control instructions to related equipment for execution. According to the method, the probability of event occurrence can be predicted, event monitoring is converted from pure passive response to active pre-judgment, risk early warning is carried out, key disposal factors can be identified, disposal parameters can be quantified, and therefore a better disposal scheme is formed.

The invention provides a self-adaptive industry report generation system based on an evolutionary cognitive model, and belongs to the technical field of artificial intelligence and industry analysis crossing. According to the method, end-to-end automatic generation and continuous self-optimization of the industrial analysis report are realized through multi-source data access and standardization, time sequence feature modeling and trend identification, evolutionary cognitive modeling and semantic skeleton generation, dynamic generation and content alignment, and adaptive feedback and cognitive evolution circulation. Compared with a traditional method, the method has the remarkable advantages in the aspects of solving the problems of logic chain breakage, knowledge updating lag, insufficient specialty and adaptability and the like in report generation. By constructing a cognitive model with continuous learning ability and combining reinforcement learning and a multi-objective optimization mechanism, the system can simulate expert thinking and produce an industrial analysis report with deep insight and high adaptability.

The invention discloses a digital textbook editing and self-adaptive learning method and system based on artificial intelligence, and relates to the technical field of education informationization, and the method comprises the steps: automatically adding a structured tag representing teaching semantics for a content element; semantic analysis is carried out based on the structured label, knowledge entities and relationships are extracted to dynamically update an education knowledge graph, and the knowledge state of the learner is tracked; responding to the learning event, synthesizing the education knowledge graph, the knowledge state and the personalized cognitive model, and generating a personalized adaptive learning path; and scheduling and presenting adaptive teaching resources through an edge intelligent distribution network based on the path. According to the invention, a full-link closed loop from content intelligent creation, knowledge systematic construction, learning path personalized planning to resource efficient delivery is realized, the technical problems of low editing efficiency, single content, lack of personalization, data isolated island and the like of an existing digital textbook system are effectively solved, and the teaching and learning effects are remarkably improved.

The invention belongs to the technical field of industrial automation and intelligent operation and maintenance, and relates to a belt conveyor multi-mode operation and maintenance fusion system and method based on a three-tower type cognitive model. The method comprises the following steps: carrying out feature extraction on multi-modal operation and maintenance data of a belt conveyor, and uniformly mapping the extracted features to the same vector space to obtain multi-modal joint feature vectors with consistent dimensions; performing feature fusion processing on the multi-modal joint feature vector to obtain a multi-modal cognitive feature vector for representing the overall running state of the belt conveyor; according to the current operation and maintenance task type, activating the expert sub-network corresponding to the current operation and maintenance task type, and outputting a risk vector based on the multi-modal cognitive feature vector; executing comprehensive analysis and situation reasoning based on the risk vector to obtain a belt conveyor operation situation evaluation report and a corresponding workability score; and performing causal reasoning based on the multi-modal cognitive feature vector and the risk vector to obtain an optimal operation and maintenance decision matched with the current operation state.

The application belongs to the technical field of industrial automation and intelligent operation and maintenance, and relates to a belt conveyor multi-modal operation and maintenance fusion system and method based on a three-tower cognitive model. The method comprises the following steps: feature extraction is performed on multi-modal operation and maintenance data of the belt conveyor, and the extracted features are uniformly mapped to the same vector space to obtain multi-modal joint feature vectors with consistent dimensions; feature fusion processing is performed on the multi-modal joint feature vectors to obtain multi-modal cognitive feature vectors for representing the overall operation state of the belt conveyor; according to the current operation and maintenance task type, an expert subnetwork corresponding to the task type is activated, and a risk vector is output based on the multi-modal cognitive feature vectors; comprehensive analysis and situation reasoning are performed based on the risk vector to obtain a belt conveyor operation situation evaluation report and a corresponding workability score; causal reasoning is performed based on the multi-modal cognitive feature vectors and the risk vector to obtain an optimal operation and maintenance decision matched with the current operation state.