259 results about "Context sensing" patented technology

The invention provides a multi-task printing scheduling optimization method based on AI intelligent strategy generation, and relates to the technical field of production workshop intelligent scheduling, and the method constructs and maintains a dynamic cognitive map, comprehensively perceives each entity state and complex association relationship of printing production in real time, utilizes an artificial intelligence engine, carries out context perception and reasoning based on the map, and improves the printing quality. The system executes the process in real time, analyzes event data, incrementally updates the state attribute of the cognitive map, evaluates the current execution effect, and adjusts the execution path and resource allocation of the flexible workflow in real time by applying a learning strategy; and continuous self-optimization and system self-evolution of a scheduling strategy model are realized through historical execution data, so that the problems of response rigidity, difficulty in realizing multi-target collaborative optimization and lack of fine-grained real-time adjustment in coping with emergencies in traditional scheduling are solved. And the scheduling flexibility, the overall efficiency, the product quality and the capability of coping with a complex and uncertain environment of printing production are remarkably improved.

The invention relates to the technical field of semantic recognition, can be applied to business scenes such as financial science and technology and medical health, and discloses an interactive code generation method, device and equipment and a medium. And a visual preview interface is generated in combination with the context, multi-round modification through interactive adjustment or text editing is supported, and finally user interface element codes are generated. According to the method, natural language processing, context sensing and visual interactive correction are combined, so that the closed-loop process of interface codes from generation to dynamic adjustment is realized, the customization degree and the service adaptability of a generated result are improved, and the problems that an existing tool template is rigid, semantic understanding is deviated, and a multi-round correction mechanism is lacked are solved.

The invention relates to the technical field of intelligent retrieval, in particular to an enterprise RAG intelligent question answering system based on vector database reordering. The system specifically comprises: a document recall module, which retrieves a vector database to obtain candidate document blocks containing business metadata; the comprehensive scoring module is used for calculating a semantic correlation score by adopting a later-stage interaction architecture based on bidirectional token importance weighting, performing path semantic matching and context sensing rule evaluation according to a preset metadata ontology graph to obtain a service attribute score, and analyzing the evidence sub-graph to obtain a fact path score; fusing the semantic correlation score, the service attribute score and the fact path score to generate a comprehensive correlation score; and the sorting output module performs optimization resorting based on a preset punishment mechanism and the comprehensive correlation score to generate an optimized context set, and calls a generation model to output answers based on the optimized context set. According to the method, semantic accuracy, business compliance and fact reliability can be considered, and more trustworthy high-quality enterprise-level answers can be generated.

An auditory operating system designed to facilitate context-aware, audio-based user interactions, particularly with artificial intelligence agents or applications. An auditory operating system shell serves as the primary interface, managing and coordinating multiple specialized agents that handle specific domains like music streaming, scheduling, or home automation. Using natural language processing, the auditory operating system shell identifies the appropriate agent or application for a user's command or query and ensures task execution and context preservation across interactions. The auditory operating system shell enforces privacy and stability by controlling agents' and applications' access to data and system privileges. The auditory operating system shell also supports dynamic context management, enabling seamless handoffs between agents when user requests span multiple domains. This auditory operating system shell may reduce the need for users to memorize specific wake words or commands, as the auditory operating system shell may determine the user's intent from speech or contextual cues.

A system for managing financial portfolio as well as for recommending personalized investment strategies, is disclosed. The system includes a first computing unit having an application interface, communicably connected to a central controller. The central controller includes a back-end server. The backend server includes a data receiving component adapted to receive the input data-sets from the first computing unit and real time data-sets from a plurality of data-sources. The backend server further includes a data analysis module adapted to process the real-time data to generate one or more actionable insights. The backend server furthermore includes a contextually intelligent portfolio management module adapted to utilize one or more contextual data related to the user, to monitors the user's investments and asset portfolios, and generate contextually relevant portfolio information for the user in a real-time. The backend server additionally includes a financial strategy implementation module adapted to utilize the actionable insights in combination with the contextually relevant portfolio information of the user to generate personalized investment strategies and recommendations for each user. In operation, a user generates and / or formulate at least one input query based at least in part on one or more input data-sets related to the user's financial portfolio. Thereafter, the input datasets are received at the back-end server which in turn are processed by the financial strategy implementation module in combination with one or more actionable insights generated by the data analysis module of the back-end server to identify a response to the input query, and / or provide one or more personalized investment related recommendation. The identified information and / or recommendation is elicited as a response to the input query and is presented and / or visualized on an output component of the first computing unit.

The invention discloses a multi-source heterogeneous knowledge fusion question and answer solution system, which relates to the technical field of computers and comprises a knowledge base layer, a recall layer, an analysis and pruning layer and an answer generation layer. The knowledge base layer is used for constructing a dynamic heterogeneous entity fusion engine and a self-adaptive knowledge slice storage mechanism, and the dynamic heterogeneous entity fusion engine comprises a cross-modal entity alignment algorithm, relation topology completion and an incremental entity evolution model; the self-adaptive knowledge slice storage mechanism is used for performing semantic density perception slicing on the RAG document and constructing a three-level index tree; according to the method, through the technologies of cross-modal entity alignment, dynamic recall weight adjustment, context sensing entity priority correction, RAG cross-block reasoning enhancement, double-encoder conflict detection, knowledge graph guide generation and the like, knowledge fusion reasoning is achieved, the complex question answering accuracy is effectively improved, the conflict recognition rate is increased, and the manual maintenance cost is reduced.

The invention relates to the technical field of computers, and provides a multi-modal learning resource intelligent recommendation method and system based on an AI large model, and the method comprises the steps: carrying out the intention analysis of interaction behavior data and user information based on the AI large model, and obtaining a user feature vector; obtaining multi-modal resource data based on the user feature vector, and integrating the multi-modal resource data to obtain a resource representation vector; mining implicit association between resources and knowledge points and a mapping relation between user requirements and target skills based on a knowledge graph in combination with user feature vectors and resource representation vectors to obtain a knowledge network; generating candidate recommended learning resources based on the context awareness information and the knowledge network; and performing sorting optimization on the candidate recommendation learning resources based on the user feature vector and the resource representation vector in combination with the current stage index of the multi-modal resource data, and generating an initial resource recommendation list. According to the embodiment of the invention, the content features of the learning resources are comprehensively captured, and the resource recommendation accuracy is improved.

The invention relates to the technical field of medical image processing, and discloses a method for intelligently describing liver space-occupying lesion ultrasonic image content by using LLM. A liver ultrasonic image sequence, a patient historical medical record text and a blood biochemical index vector are obtained through a multi-modal data acquisition module, and features are extracted through a cross-modal contrast learning network to generate embedded vectors and align the embedded vectors. And inputting the aligned image embedding vector into a dynamic context sensing decoder, and generating a description text semantic mark sequence by using a layered multi-head attention mechanism. The confidence coefficient is evaluated through an uncertainty calibration module, and the text is optimized through a post-processing reordering mechanism when the confidence coefficient is lower than a threshold value. A real-time interaction optimization mechanism is further arranged, and the model is updated according to feedback of doctors. According to the method, multi-modal data are fused, description accuracy and reliability are improved, text quality is optimized, clinical requirements are met, and liver disease diagnosis is assisted.

The invention discloses a personalized computer-aided decision-making method and system fusing multi-modal data, and relates to the field of personalized computer-aided decision-making, and the method comprises the steps: mapping multi-source heterogeneous modal data to a unified semantic embedding space, and obtaining a multi-modal unified representation vector set; carrying out three-layer progressive fusion on a feature layer, a situation layer and a decision layer of the multi-modal data to generate a global decision context vector; based on a cross attention mechanism, outputting a fused context sensing personalized vector; based on the behavior cloning model, outputting probability distribution on all decision options; according to the user feedback operation data, generating a user personalized decision strategy and performing dynamic optimization; and generating a structured decision report containing visual traceability information based on the hierarchical fusion process and decision reasoning logic. End-to-end intelligent generation from multi-source heterogeneous data to personalized decisions is realized, and a standardized process is converted into personalized customized decisions.

A decision support system and a computer-implemented method of enterprise decision support include a data consolidation module, domain-specific machine learning models, an enterprise decision intelligence model, and a user interface layer. The data consolidation module collects data from both internal and external data sources. The domain-specific machine learning models generate decisions based on the collected data. The enterprise decision intelligence model integrates real-time trends and the decisions to provide context-aware recommendations. The enterprise decision intelligence model maintains a decision graph that connects decision variables of the domain-specific machine learning models in a causal relationship, with the domain-specific machine learning models interacting as an interconnected network. The decisions are influenced by affects of decision variables from other domain-specific machine learning models. The user interface layer facilitates interactive decision-making by way of the enterprise decision intelligence model and visualizing the recommendations.

The invention discloses a sleep staging method and system based on multimode signal fusion and context awareness. The sleep staging method based on multimode signal fusion and context sensing comprises the following steps: acquiring a multimode physiological data set; performing multi-scale feature extraction on the multi-modal physiological data set to obtain a multi-scale feature set; fusing the multi-scale feature set based on a lightweight model to obtain a fused feature set; classifying the fused feature set, and obtaining a sleep staging result based on a classification result; wherein the multi-scale feature extraction of the multi-modal physiological data set comprises the following steps: segmenting the multi-modal physiological data set into a frame sequence set; and performing historical context feature extraction, future context feature extraction simulation and timestamp feature extraction on the frame sequence set to obtain a context feature set. In this way, the staging continuity can be enhanced, and the misjudgment rate caused by burst signal fluctuation is reduced.

A system for filtering and presenting notifications received by an aircraft includes a processor, a display device, a communication device, and computer-readable memory. The computer-readable memory encoded with instructions that, when executed by the processor, cause the system to perform the following steps. The system receives one or more broadcast notifications. The notifications are, for example, Notices to Air Missions (NOTAMs). The system filters the one or more broadcasted notifications based upon applicability to a present flight plan to generate applicable notifications. The system filters the applicable notifications based upon one or more filtering criteria to assign a criticality to each of the applicable notifications. The system display, via the display device, the applicable notifications based upon the criticality of each of the applicable notifications. The system outputs alternate flight plans due to restrictions on the present flight plan.

The invention relates to the technical field of data analysis, can be applied to business system platforms of financial science and technology, medical health and the like, and discloses a user intention recognition method, device and equipment and a medium. Generating an intention recognition strategy of each user intention point according to the key intention verbal skill, constructing an intention recognition model, carrying out context semantic analysis on interaction text data of a target user, generating a context perception text, carrying out distributed coding on the key intention verbal skill in the intention recognition strategy, and obtaining intention strategy representation; and determining an attention weight according to the context sensing text and the intention strategy representation, identifying intention key data of the interactive text data, performing intention analysis on the intention key data by using an intention identification model, and generating a target intention of the target user. According to the invention, the accuracy and efficiency of intention recognition of the target user are improved.

The invention provides an equipment operation state anomaly detection method based on edge computing. According to the method, the accuracy, the robustness and the real-time response capability of equipment anomaly detection are improved. The method comprises the following steps: firstly, setting a channel confidence coefficient screening mechanism at an edge calculation end, and eliminating an abnormal transmission channel with drifting or fluctuation to guarantee data quality; then, fusing and modeling a structural dependency relationship among multiple channels by utilizing a graph neural network and a self-attention mechanism, and extracting a global context sensing vector; and finally, constructing a dynamic criterion system based on a multi-task state index, and realizing efficient anomaly judgment in combination with state difference scoring. According to the invention, through the integration of edge-side high-quality data screening, depth feature modeling and a fusion type detection mechanism, the intelligent level and field adaptability of equipment operation state abnormity identification are significantly improved.

An electromechanical equipment knowledge graph link prediction method fused with a large language model belongs to the field of electromechanical equipment and knowledge graphs, and comprises the following steps: step 1, constructing an electromechanical equipment knowledge graph by combining a fine tuning large language model and a prompt project; step 2, designing an encoder with a multilayer graph attention network for the constructed EKG, and realizing embedded representation updating of the knowledge graph; and a third step, combining a decoder to decode the knowledge graph embedding obtained by coding, and executing a link prediction task. According to the method, effective extraction of the triple information of the high-quality electromechanical equipment is realized, and a data basis is provided for fault root cause analysis and fault prediction of the electromechanical equipment; the representation capability of the graph is enhanced, the context sensing capability of the knowledge graph is enhanced, and the adaptability and prediction performance of the model in the field of electromechanical equipment are improved.

A document-level intelligent manufacturing process flow relation extraction method comprises the following steps: S1, acquiring a process document, and labeling process entities in the process document and a relation between the process entities; s2, performing deep coding on the process document by using a pre-training language model, extracting core semantic information of the process document, and generating process entity representation with consistent semantics based on an entity aggregation strategy of context sensing; s3, nodes and edges of a heterogeneous graph are constructed according to process entity representation, and a relation graph convolutional network R-GCN and a hierarchical attention mechanism are introduced to fully capture the relation between process entities; according to the method, the modeling capability of technological process sequence constraint and the fusion effect of document semantics and graph structure features are respectively enhanced by a sequential construction strategy and a process perception double-graph distillation mechanism oriented to a manufacturing process, the integrity and the utilization rate of information transmission are improved, the performance loss caused by error layer-by-layer propagation is relieved, and the method is suitable for popularization and application. Accurate and efficient extraction of the entity relationship in the process document is realized.

The invention provides a YOLOv7-based photovoltaic module fault detection method and a YOLOv7-based photovoltaic module fault detection device, and relates to the technical field of target detection. According to the method, a thermal infrared and temperature information fusion module is introduced at the front end of a network, a thermal infrared image and a pixel-level temperature matrix thereof are fused through alignment, coding and attention mechanisms, and the characterization capability of multi-modal features is enhanced. A shallow layer-deep layer information aggregation module is introduced into the neck network, and the context perception and feature discrimination ability of the model to a multi-scale small target is improved. A global context sensing module is introduced in front of a detection head, a directional channel context branch and a query-key value branch of the global context sensing module capture directional global dependence and spatial context relations respectively, and global modulation of features is achieved. According to the method, the detection precision and robustness of various faults such as faults, fragmentation, hot spots and shielding of the photovoltaic module junction box are remarkably improved, the weak and small target positioning capability is optimized, the method is suitable for being deployed on mobile or embedded equipment, and efficient and intelligent inspection of a photovoltaic power station is achieved.

The invention discloses an image segmentation method and system fusing sequence modeling and a multi-scale attention mechanism, and relates to the field of computer vision, and the method comprises the steps: carrying out the extraction of multi-scale detail and space information through an encoder based on a multi-scale convolution residual fusion module, and gradually carrying out the downsampling to increase a receptive field; performing multi-scale and global context feature extraction and coding based on a multi-scale context global modeling module through a bridging layer; and through a decoder, encoder features are fused based on up-sampling, convolution and a double-guide gating fusion module, so that recovery and refinement of high-resolution semantic features are realized, and an image segmentation result of fusion sequence modeling and a multi-scale attention mechanism is obtained. According to the method, the multi-scale features are fully extracted and fused, the context sensing capability is enhanced by using sequence modeling, the features of the encoder and the decoder are dynamically fused through a gating mechanism, the segmentation precision and the edge detail performance are improved, and the problem that the model performance and generalization capability are limited in the prior art is solved.

The invention discloses a family education auxiliary method and interaction system based on an intelligent agent mode, and belongs to the technical field of data processing, and the method specifically comprises the steps: obtaining the multi-source data of family education, building an intelligent agent model of cross-dimension situation perception capability based on the multi-source data, and enabling the intelligent agent model to carry out the multi-dimension situation perception capability according to the evolution trend of three-dimensional association, the intelligent agent model automatically deduces the appearing negative education situation, when a risk node is predicted, an education intervention scheme is generated, the education intervention scheme comprises a learning task and a real-time prompt for a parent communication mode, and in the execution process of the education intervention scheme, the intelligent agent model carries out different levels of task pushing on students and parents at the same time, so that the education intervention effect is improved. The student side obtains learning tasks, the parent side obtains interactive guidance, cooperative auxiliary optimization is carried out on family education, learning progress and parent-child communication can be promoted in the same education process, and it is ensured that the education process has dynamic adaptability and systematic management ability.

The invention discloses a radiation safety management method and system based on cloud platform data driving, and relates to the technical field of cloud platform radiation management. The method comprises the following steps: collecting radiation field data and environment state data of a target area through a deployed intelligent sensing node; uploading the radiation field data, the environment state data and the context data from the service module to a cloud platform, performing energy compensation and radiation unmixing, and constructing a multi-dimensional feature vector; and processing the multi-dimensional feature vector by using a pre-trained situational radiation perception model, identifying radiation field features and an environment situation, and generating a situational radiation safety early warning signal. The technical problem that in the prior art, radiation safety monitoring depends on single-point measurement, comprehensive judgment cannot be carried out in combination with environment and service context data, and consequently the radiation field anomaly recognition capability is insufficient is solved, and the purpose that the radiation field anomaly recognition capability is improved through cloud platform data driving and context awareness model fusion is achieved. And the technical effects of high-precision identification and situational safety early warning of the radiation field state are realized.

The invention discloses an SAR ship instance segmentation method for self-adaptive representation alignment, and belongs to the field of SAR image instance segmentation. According to the implementation method, the problems of semantic-structural feature mismatch, global-local feature extraction mismatch and cross-dataset scale generalization mismatch are solved through modular design by constructing an adaptive representation alignment network. The method comprises an edge-guided boundary optimization module, a context sensing module and a depth adaptive feature pyramid module. The boundary optimization module introduces edge enhancement information in the feature extraction process to improve the target boundary positioning precision; the context sensing module fuses a multi-path global attention mechanism in a deep feature stage, and the semantic discrimination capability is enhanced; the depth adaptive feature pyramid module adaptively selects the optimal feature fusion depth by constructing a multi-depth fusion path in combination with scale statistical information, and improves the cross-dataset robustness. According to the method, the segmentation precision of the SAR ship instance can be remarkably improved.

The invention discloses a synthetic aperture radar image target detection method based on a multi-scale residual fusion Transform and a space-channel attention mechanism. The method comprises the following steps: S1, generating a synthetic aperture radar image data set by using a synthetic aperture radar image target detection system; s2, inputting the synthetic aperture radar image data set into the structure of a deep neural network for training to obtain a trained deep neural network; and S3, inputting a to-be-detected synthetic aperture radar image into the trained deep neural network to obtain a synthetic aperture radar image target detection result. According to the method, a space-channel attention module is used, a double-branch attention mechanism is adopted, collaborative expression of global context and local details is enhanced, and context sensing capacity is enhanced to suppress background interference. According to the method, a residual fusion Transform module is used, and a residual fusion and feature pyramid injection structure is adopted, so that feature expression is enhanced.

The invention discloses a corn field weed detection method and device based on improved YOLOv11. The method comprises the following steps: acquiring and constructing a weed image data set for model training; replacing a backbone network of the YOLOv11n baseline network with a double-flow visual network, and taking the replaced network as a first optimized network; replacing the original pyramid pooling module with an enhanced receptive field module at the tail end of the backbone network of the first optimized network to obtain a second optimized network; in a neck feature fusion layer of the second optimization network, a self-adaptive context guide fusion module is used for replacing traditional splicing, and a third optimization network is obtained; and taking the third optimization network as a final lightweight high-precision weed detection model DEA-YOLO11, and detecting field weeds based on the model. According to the method, the problem of low detection precision caused by insufficient global context sensing capability, fine-grained feature loss and low multi-scale feature fusion efficiency of a weed detection model in a complex agricultural scene in the prior art is solved.

The invention provides an urban traffic management intelligent evaluation system and method based on a large language model, and belongs to the technical field of intelligent traffic management. The system comprises a data processing module, an MECA traffic flow analysis module, a feature engineering module, a driving behavior evaluation module, a clustering analysis module, an LLM intelligent decision module, a visualization generation module and a visualization module. The system analyzes the road traffic data collected by the unmanned aerial vehicle, adopts the MECA technology to automatically identify the road type and the traffic environment, intelligently judges the congestion level, evaluates the driving behavior, and combines a big language model to generate a targeted traffic management optimization suggestion. According to the method, the adaptive context sensing technology is innovatively combined with multi-criterion learning, adaptive congestion judgment of different road types is realized, different traffic characteristics of urban expressways, common urban roads and expressways can be accurately recognized, and differentiated management strategies are provided accordingly. The system supports real-time processing of large-scale traffic data, and provides scientific and accurate decision support for urban traffic management departments.

The invention discloses an intelligent agent planning method, system and device based on context awareness and a medium, and belongs to the technical field of artificial intelligence. The method comprises the steps that firstly, user input is analyzed, structured task requirements are generated, and meanwhile, external tool function attributes and historical interaction data are collected; according to the method, an ordered execution plan is generated in combination with task requirements and tool attributes, subsequent plans are optimized by utilizing historical interaction contexts, in an execution stage, parameters required by a target function can be accurately extracted from the task requirements and context information, corresponding tools are called to complete a task, and finally, an execution result is fed back for display, so that the execution efficiency is improved. And performing execution control based on the function calling state and the result. According to the method, more intelligent task planning and execution are realized through context sensing, and the method is particularly suitable for complex task scenes needing multi-step and multi-tool cooperation.

The invention relates to the technical field of language model integration, in particular to a large language model integration method supporting semantic correction. The method comprises the following steps: S1, selecting several large language models, and sorting and preprocessing text data for training and testing; s2, training each source model and an optimization model by using an intelligent fusion language model technology; s3, fusing parameters of each source model by using an intelligent fusion language model technology, selecting model parameters for fusion according to the text data, and adjusting the fused model; s4, evaluating the semantic correction capability of the fused model, and adjusting a fusion strategy and regularization parameters according to an evaluation result; and S5, integrating the trained model into a target system. According to the large language model integration method supporting semantic correction, through an intelligent fusion language model technology, a context sensing and hybrid regularization technology is utilized to train and optimize a model and enhance adaptability, and a FuseLLM fusion technology is adopted to merge all source model parameters.

The invention discloses a termite remote identification method based on image processing, and aims to solve the problem that termite activities and out-of-class interference are difficult to reliably distinguish only by indirect visual clues. According to the method, visual representation of an image sequence is extracted through a context awareness framework; utilizing a time sequence diffusion model to analyze a time evolution mode of the indirect visual clues, and evaluating whether the indirect visual clues accord with long-term and continuous activity characteristics of the termites or not; aiming at the uncertain region, inferring the form or evolution trajectory of the uncertain region in combination with a conditional diffusion model, evaluating the diagnosis reasonability, and quantifying the similarity between the observation phenomenon and typical termite activities and extraclass interferents; integrating a preset knowledge graph, and introducing environment context information to carry out matching verification; a plurality of dimensions such as evolution trend, diagnosis rationality, out-of-class interference similarity and context matching degree are integrated, and termite activities are accurately identified through a situation-adaptive fusion decision model; according to the method, the reliability of remote non-destructive monitoring is effectively improved.

The invention provides a UI automatic test method and test system and a storage medium. The method comprises the steps that an optical character recognition result of a target text and a to-be-tested UI is obtained, and the optical character recognition result comprises a plurality of elements and element texts, coordinate frames and visual features of all the elements; obtaining the semantic similarity between the target text and each element text; obtaining attribute similarity between the target text and each element based on the visual features of each element; according to the semantic similarity and the attribute similarity, elements, corresponding to the target text, in the optical character recognition result are determined to serve as target elements; taking the position of the coordinate frame of the target element in the UI as a target position; and generating a presentation image which highlights the target position and / or the target element. Therefore, fuzzy matching based on semantics can be realized, multi-modal features including text semantics and visual features are fused, and the robustness and accuracy of UI element positioning are improved in combination with context sensing.

The invention belongs to the technical field of intelligent interaction, particularly relates to an intelligent interaction method, system and device based on AI communication and a storage medium, and aims to solve the problems that in the prior art, AI scene coverage is narrow and is separated from a business process; the system comprises four modules: a context sensing and modeling module which collects multi-source context information in real time, and generates a three-dimensional context tensor through structured alignment and time sequence normalization; the intention analysis and reasoning module performs intention level decomposition by using a deep semantic model based on the tensor, and outputs an intention recognition result and probability distribution; the response generation and adaptation module is used for generating multi-mode responses such as texts, voices, images or equipment instructions in combination with intention and context tensors, and outputting the multi-mode responses after consistency verification; and the interactive feedback and optimization module collects user explicit and implicit feedback and is used for updating model parameters and strategy weights online to realize continuous optimization of the system.

The invention discloses a low-voltage series arc fault detection method, system and device, and belongs to the technical field of low-voltage series arc fault detection, and the method comprises the steps: obtaining an original current signal, and carrying out the preprocessing through sliding window segmentation and instance normalization; performing multi-scale feature fusion on the preprocessed analysis unit, and generating fusion features through parallel feature extraction and an attention mechanism; performing context modeling on the fused features through an encoder, inputting a self-adaptive bottleneck layer containing an expert hybrid network, and routing the features to the most appropriate expert network by context sensing gating according to global information; the decoder reconstructs the signal and calculates an error, and generates a dense abnormal fraction sequence; gaussian position weighted aggregation abnormal scores are adopted, and fault judgment is carried out in combination with a self-adaptive threshold decision mechanism based on K-Means clustering. The method can be trained without a fault sample, can dynamically adapt to complex current modes under different loads, gets rid of dependence on the fault sample, and accurately detects the arc fault.