209 results about "Information space" patented technology

An intelligent analysis method based on medical document structure perception and multi-modal fusion comprises the following steps: carrying out structure topology modeling on a medical document, extracting visual layout, text meta-information, space coordinates and semantic keyword features, constructing a semantic topological graph and dynamically shielding irrelevant contents; selecting an extraction path according to a document type, performing deep semantic analysis and entity recognition on a text-type document, and performing visual enhancement OCR recognition on a scanning-type document; the features are injected into a medical knowledge graph, and feature fusion, semantic verification, relation reasoning and information completion are achieved through a graph neural network; a three-stage strategy optimization model of basic pre-training, domain adaptation and online reinforcement learning is adopted; and large-scale processing is realized through a dynamically aggregated distributed architecture. The method is used for intelligent analysis and structured conversion of documents of hospitals, medical insurance and medical scientific research. The problems that heterogeneous medical document analysis adaptability is poor, multi-modal fusion is difficult, medical knowledge utilization is insufficient, and large-scale processing efficiency is low are solved.

A method for electronic processing of user queries maintaining factual consistency during processing includes receiving a user query, executing a hybrid retrieval operation, generating a context information space that includes a set of paragraph output from the paragraph retrieval operation, a set of triplet output from the triplet retrieval operation, and a set of graphs output from the graph retrieval operation. Furthermore, generating a first prompt instruction to generate an intermediate rationale as explanation output and performing an automatic fact-checking operation to verify the explanation output and re-feeding the user query along with the generated explanation output to the generative language model to derive an answer output that is succinct in one or more language parameters as compared to the generated explanation output. The answer output includes a verified and evidence-supported answer and at least one fact reference from first set of factual information with reduced hallucinations and factual inconsistencies.

The invention discloses a patent disclosure auxiliary writing method and system based on multi-agent collaboration. The method comprises the steps that a multi-agent system composed of a collaboration agent, an agent agent and an expert agent is constructed, a knowledge base and a decision module of each agent are configured, and a collaboration architecture of a task scheduler, an arbiter and a shared information space is established; the method comprises the following steps: receiving an original technical scheme of a user, and constructing a structured intermediate representation through multi-agent parallel semantic analysis and feature extraction; based on the representation, a writing task sequence is generated through dynamic task decomposition, intelligent agent division is coordinated through a contract network protocol, and content generation and verification are completed based on a multi-round debate mechanism; performing multi-dimensional consistency detection on the generated draft, and eliminating logic conflicts and expressing defects through iterative revision; and finally, an interest portrait is generated based on the user technical scheme and the optimized draft, interest retrieval is performed, a technical content report is output, and the efficiency and quality of disclosure book writing are improved.

The invention relates to an offshore wind power operation environment scene construction method considering typhoon forecast errors. The method comprises the steps of selecting similar historical typhoon forecast data according to current typhoon forecast information; based on wind speeds of similar historical typhoon forecast and current typhoon forecast, combining marine environment element information to construct a marine environment scene information space; a kernel density estimation model of historical typhoon wind speed forecast errors is constructed, corresponding kernel density bandwidths are calculated and updated according to different typhoon forecast timeliness, and historical forecast wind speed error distribution is obtained; and inputting the marine environment scene information space and the historical forecast wind speed error distribution into the improved generative adversarial network, and generating a marine environment scene of which the wind speed interval meets the historical typhoon wind speed forecast error distribution through the wind speed interval constraint loss. Compared with the prior art, the marine environment scene is updated according to the wind speed error distribution under different forecast time periods of the typhoon, and the scene generation quality is improved.

The invention provides a station group runoff prediction method based on a depth time sequence graph network, which is characterized in that a new station group runoff prediction model based on a graph neural network and a convolutional neural network is constructed for runoff prediction of multiple stations in a region, and a time sequence information extraction module is used for extracting feature information of input data on a time scale; the spatial aggregation module is used for aggregating feature information among different sites and describing the influence of other sites on the site; and the time sequence modeling part is responsible for performing time sequence modeling on information obtained by aggregation of each station, and finally performing full network connection on hidden state information to obtain predicted runoff of each station. According to the method, geographic space relations and mutual influence factors existing between different stations can be efficiently extracted, meanwhile, a traditional sequential network is fused, space-time relevance and consistency are comprehensively considered, hidden space-time relations in data are fully mined, and the accuracy of overall runoff prediction of the station group is improved.

The invention discloses a weak light image enhancement method and device, equipment and a storage medium, and relates to the technical field of image enhancement, and the method comprises the steps: obtaining a to-be-enhanced image and a target normal illumination image, and optimizing an initial image enhancement model to obtain a target image enhancement model; performing feature extraction and manifold distribution estimation on the target normal illumination image by using the initial image enhancement model to obtain first distribution manifold information; performing feature extraction and manifold distribution estimation on the to-be-enhanced image by using the target image enhancement model to obtain second distribution estimation information; fusing the first distribution manifold information and the second distribution estimation information by using a target image enhancement model, and establishing a mapping relation between target high-dimensional feature information vectors and feature information vectors by using the target image enhancement model based on a fused target high-dimensional manifold information space; and obtaining a target projection and target manifold distribution according to the mapping relationship, and decoding the target manifold distribution by using a decoder to obtain a target image enhancement result. And image adaptive enhancement is realized.

A transmission control device and method for BIM (Building Information Modeling) information, comprising the following steps: analyzing the BIM information of a source building to initially obtain a sub-period cluster of the BIM information of the source building so as to obtain a plurality of information spaces according to the change of information in the sub-period cluster; obtaining clutter value information space and non-clutter value information space according to the difference of information variation between the information spaces, and obtaining the importance of each clutter value information space through the difference of real clutter value levels between the clutter value information space and the non-clutter value information space; and performing clutter value removal on the source building BIM information through a Wiener filtering method according to the obtained importance degree. According to the method, the building BIM information with a plurality of information sources can be distinguished, so that the clutter removal value of the source building BIM information is more complete, and the effective building BIM information contained in the source building BIM information can be clearly identified.

The invention discloses a zero-trust API dynamic access control method, a computer device and a medium. The method comprises the following steps: processing all collected original identity information of a user in different systems through a preset aggregation strategy, and generating a unified decentralized digital identity label of the user based on a block chain; a basic governance structure of DAO is configured on a block chain network, and based on the structure, an initial access strategy corresponding to each user is determined in a DAO member voting mode; acquiring multi-dimensional information for initiating access, and performing multi-dimensional risk assessment on the multi-dimensional information including time dimension information, space dimension information, equipment dimension information and network dimension information based on a preset multi-dimensional weight to obtain comprehensive risk assessment information; and obtaining an access control strategy according to the comprehensive risk assessment information and the initial access strategy. The method can adapt to rapid identity authentication in different service systems and security environments, and is high in flexibility and security.

The invention relates to the technical field of intelligent operation and maintenance of mechanical equipment, in particular to an electromechanical composite transmission system self-lifting diagnosis method based on a memory sub-network, and the method comprises the steps: constructing a self-lifting fault diagnosis framework which comprises a health category prediction network, a stable information memory sub-network and a rapid information memory sub-network; in the autonomous lifting learning stage, a fault diagnosis data set A is randomly extracted, and a data set B is randomly extracted from a learned information space; predicting the data set B by using the two sub-networks, and selecting an optimal playback log; predicting the data set A and the data set B on the basis of a health category prediction network to obtain two logs; calculating the total loss based on the playback logits and the two logits of the health category prediction network, and updating the weight and the learned information space of each network; and carrying out multiple times of autonomous lifting learning until a training target is met. According to the invention, the fault diagnosis network can be continuously trained from the continuously added data streams, and the autonomous improvement of the network is realized.

The invention discloses a thyroid nodule benign and malignant prediction method based on automatic segmentation and semi-supervised learning. The thyroid nodule benign and malignant prediction method comprises the following steps of data set establishment, image preprocessing, thyroid nodule segmentation model training, mask and image fusion processing, thyroid nodule benign and malignant classification model training and semi-supervised learning model construction. According to the method, the thyroid nodule effective area is segmented firstly, subsequent benign and malignant classification tasks are assisted, and information space is expanded and learned through semi-supervised learning, so that the problems of high medical image marking cost, insufficient data and limitation on deep learning model performance in the prior art are solved; and meanwhile, the second-generation basic model of the sliding window layered visual converter after mixing fusion is carried out on the basis of pre-training and by using the nodule region generated by the thyroid segmentation model realizes good performance, and high-precision thyroid nodule benign and malignant classification prediction can be realized.

The invention discloses an intelligent building management analysis system based on an AI large model, and relates to the technical field of management analysis systems.The intelligent building management analysis system comprises a statistics module, an analysis module, a link module and a function module, the statistics module is used for collecting virtual information, space information and user information in an intelligent building; the link module is used for performing interface API adjustment and data transmission path regulation and control on the management analysis system of the intelligent building; the function module is used by a user for direct application of the intelligent equipment and overall management of the management analysis system; according to the intelligent building management analysis system, seamless mutual linkage and data transmission between the intelligent building management analysis system and other different AI large models can be achieved through the conversion interface, a user can select the needed AI model to operate and use all intelligent devices in the intelligent building where the user is located, and multiple requirements of modern city development are met.

A video compression system and method integrates geometric compression with cognitive understanding through a persistent cognitive machine interface. The system employs a hierarchical encoder generating multi-scale compressed representations organized within a Lorentzian manifold structure. A geometric processor maintains temporal causality through time-like geodesics and light cone constraints while organizing video content according to semantic relationships. A cognitive interface creates thought bundles as navigable submanifolds, enabling semantic access to compressed content beyond traditional temporal indexing. The system supports real-time processing through progressive refinement, streaming coarse representations immediately while adding detail in parallel. Symbolic anchors mark semantically significant points, enabling concept-based navigation through compressed video. Federated learning capabilities allow distributed systems to share geometric patterns while preserving content privacy. The architecture enables improved compression ratios while maintaining both temporal causality and semantic navigability, transforming video from sequential media into an intelligently accessible information space.

The invention relates to the technical field of geographic information big data processing, and discloses a geographic information space-time big data real-time aggregation and cleaning method based on edge computing, and the method comprises the following specific steps: S1, deploying edge computing nodes in a geographic information collection terminal coverage area, constructing an edge computing network, an acquisition terminal pushes original geographic information spatio-temporal data including geographic coordinates (x, y, z), an acquisition timestamp t and attribute information to an edge node through a protocol, and the edge node establishes a data buffer queue and preliminarily classifies and stores the data according to a spatio-temporal dimension. According to the geographic information space-time big data real-time aggregation and cleaning method based on edge computing, efficient and low-delay transmission of geographic information space-time big data is achieved through edge computing node deployment and a multi-algorithm fusion aggregation cleaning strategy, and the bandwidth and delay problems of a traditional centralized architecture are solved; multi-dimensional collaborative cleaning (geographic constraint, statistics and machine learning) is combined with an intelligent algorithm, and abnormal data are accurately recognized and processed.

The invention provides a textile flexible production line multi-material scheduling system, which comprises a physical execution layer and a multi-material scheduling layer, wherein the physical execution layer comprises reconfigurable textile production equipment; the intelligent sensing layer is arranged at key nodes of the physical execution layer; the central scheduling and control layer is in communication connection with the physical execution layer and the intelligent sensing layer; and the digital twin layer is a virtual mapping of the physical execution layer in an information space, and forms a closed loop with the central scheduling and control layer and the physical execution layer. According to the method, the perspectiveness and the global optimization capability of a scheduling decision are realized through a digital twinning technology, the system places an initial production scheduling scheme in a high-fidelity virtual production line model for millisecond-level concurrent simulation deduction, and potential process bottlenecks and quality risks during blending production of different materials are accurately predicted; and by taking the production efficiency, the energy consumption and the quality qualification rate as multi-target reward functions, completing iterative optimization of massive strategies in a virtual space, thereby obtaining an approximately globally optimal scheduling instruction before the physical production is put into use.

The invention discloses a disease risk assessment method and screening device based on a multi-group student physical collaborative digital network, and relates to the field of intelligent medical detection. In order to solve the defect that multi-omics-level system collaborative analysis and robust risk assessment are difficult to realize in the prior art, the technical scheme provided by the invention is as follows: acquiring a plasma sample, acquiring a spectral signal by adopting an attenuated total reflection Fourier transform infrared spectrum, and establishing a plasma spectrum digital information space; the method comprises the following steps: constructing a biological collaborative digital network containing four nodes of protein, lipid, saccharides and nucleic acid based on pathophysiology priori knowledge, and defining node strength, edge weight and network collaborative efficiency; a health baseline configuration file is established by using a health sample, a standardized deviation score of a to-be-tested sample is calculated, a comprehensive risk score is obtained, a disease screening result is output in combination with a machine learning model, and digital evaluation of multi-omics collaborative characteristics is realized. The method is suitable for non-invasive rapid screening and risk assessment work of neurodegenerative diseases and mental diseases.

The invention discloses a multi-branch power distribution network fault positioning method and system based on multi-dimensional fault characteristics. The method comprises the following steps: acquiring fault voltage traveling wave waveform data; performing phase-mode transformation on the waveform data to extract a line-mode component, extracting wave head coordinates, amplitude and slope information of a fault traveling wave in a set time window, and after redundancy elimination, determining a feature matrix, an adjacent matrix and an edge feature tensor based on a wave head information space so as to construct a multi-dimensional fault feature space; constructing a section identification model and a precise positioning model, and performing hyper-parameter tuning on the models by adopting a parameter-setting-free harmony search algorithm; and whether fault section positioning is carried out is judged based on the multi-dimensional fault feature space, if yes, the multi-dimensional fault feature space is input into a section recognition model to obtain a section recognition result, and if not, the multi-dimensional fault feature space is input into a precise positioning model to obtain a positioning result. And dual targets of power distribution network fault section identification and accurate positioning are realized from two aspects of multi-dimensional fault feature space extraction and a power distribution network fault positioning model.

The invention provides an abnormal event positioning model training method, an abnormal event positioning method, an abnormal event positioning device and abnormal event positioning equipment. The abnormal event positioning model training method comprises the following steps: acquiring a sample monitoring image and corresponding text description information; inputting a preset chain frame reasoning template, the sample monitoring image and the text description information into the initial visual language model for reasoning to obtain multiple groups of reasoning results corresponding to the sample monitoring image, and determining format reward information of each group of reasoning results according to a preset reasoning output format and an output format of the prediction bounding box; determining space reward information of each group of reasoning results according to the prediction bounding box and the real bounding box; determining semantic reward information of each group of reasoning results according to the prediction region features and text features of the text description information; according to the format reward information, the space reward information and the semantic reward information of each group of reasoning results, parameter adjustment is carried out on the initial visual language model, and a target visual language model is obtained and used for abnormal event positioning.

The invention provides a BIM-based electromechanical pipeline intelligent collision detection and automatic optimization adjustment method, which comprises the steps of analyzing a BIM model database, and extracting component information of electromechanical components; establishing a spatial index and an adjacency relationship between the electromechanical components, constructing a component spatial relationship database, and storing the component information, the spatial index and the adjacency relationship in the component spatial relationship database; calling data in the component space relation database to carry out multi-level collision detection; performing collision optimization adjustment on the electromechanical component based on a result after multi-level collision detection, and generating an optimized BIM model to be fed back to constructors; according to the method, the optimized BIM model is subjected to 3D display, meanwhile, a figure roaming scene is constructed, the moving view angle of a constructor in a building space is simulated, collision points are secondarily checked in the roaming process, the collision recognition precision can be improved through multi-level collision detection, collision conflicts are eliminated through optimization adjustment, design errors are reduced, and the engineering efficiency is improved.

The invention relates to an efficient image super-resolution method based on a multi-branch hybrid Transformer (Multi-Branch Mixer Transformer, MBMT), in the method, a unique module of a multi-branch Token Mixer (Multi-Branch Token Mixer, MBTM) is introduced, and the module aims to replace the depth by increasing the network width, so that the learning ability and the generalization ability of the model are enhanced. The method is characterized in that the method is composed of three branches: a self-attention mechanism module, a depth convolution (Depth-Width Conversion, DWConv) and an adaptive token mixer (ATM), and each branch is responsible for capturing a long-term dependency relationship, extracting local features and expanding a global information space. According to the method, an adaptive space and a channel attention mechanism are combined, so that sufficient interaction among information extracted from the three branches is ensured, and the feature representation capability of MBTM is enhanced. Benefited from the high efficiency and the shallower network structure of the MBTM, the method provided by the invention reduces the complexity while significantly improving the SR model performance.

The invention discloses a numerical control machine tool twin model moving body general construction and expression method, and belongs to the technical field of machine tool digital twin modeling, and the method comprises the following steps: constructing a numerical control machine tool assembly geometric model, and carrying out the splitting and lightweight processing of the model; determining the motion characteristics of each part of the machine tool and the relative motion relation among the parts, and performing over-mapping processing on the geometric model and the motion characteristics; constructing a machine tool twin model topology file comprising geometric features, motion features, state features and physical motion rules; and importing the three-dimensional model of the machine tool and the topological file into a three-dimensional engine, quickly constructing a twin model mapping a physical entity of the machine tool in an information space, and carrying out visualization processing. According to the method, the geometric features, the motion features, the physical motion rules and the like of the numerical control machine tool are uniformly expressed, so that the multi-axis numerical control machine tool twin model moving body in any structural form is quickly constructed.

The invention relates to the technical field of quality detection, and discloses an HCPS-based human-computer interaction assembly quality detection method for a longitude and latitude instrument type satellite-borne laser coarse pointing mechanism, and the method comprises the steps: generating a design scheme based on design requirements in a design stage; in the manufacturing stage, the similarity of production tasks and the skill level of manufacturing personnel are evaluated, the tasks are matched with corresponding skills through strategy transfer, the AI technology is combined, diversified interaction modes are adopted, feedback of the manufacturing personnel is continuously collected, and a self-adaptive strategy of man-machine interaction is implemented; incremental learning based on events or historical data is autonomously carried out through a man-machine cooperation system; in the assembling stage, when the equipment is used for assembling the coarse pointing mechanism in the physical space, the quality of all components is checked before assembling based on the support of the information space; in the detection stage, a test work station is arranged to counteract the gravity influence of the mechanical structure, multiple parameters are measured, and the assembly quality is determined according to the corresponding detection principle. The CPM assembly quality detection based on man-machine cooperation is used for ensuring the precision.

The invention discloses a method for automatically reconstructing a complex BIM (Building Information Modeling) in a finite element platform, which relates to the technical field of finite element platforms and comprises the following steps of: acquiring geometric dimension information, spatial position information, physical parameter information and contact information of all final bodies in a model in the BIM platform; disassembling the final body to obtain Boolean operation of forming the final body by the basic body, and obtaining external calculation geometric feature information of the basic body, the process body and the final body; according to Boolean operation and external calculation geometric feature information obtained by disassembling the final body, selecting a corresponding body in the finite element platform to carry out Boolean operation, and reconstructing the final body; and selecting a corresponding final body to establish contact, and endowing the final body with corresponding physical parameter information. According to the method, geometric automatic reconstruction, contact automatic establishment and attribute automatic endowing of the BIM model in a finite element platform are realized.

The invention provides a prefabricated building component intelligent docking system based on a BIM model, and the system comprises a BIM model processing module which is used for importing the BIM model of a prefabricated building, and extracting the three-dimensional model information, geometric parameters and preset docking position information of each prefabricated component; the component sensing module is used for collecting actual position information and space attitude information of the to-be-butted component in real time; the data processing and analysis module is used for performing comparative analysis on the actual position information and the space attitude information of the to-be-docked component and preset docking position information in the BIM model, and calculating position deviation and attitude deviation; and the docking control module is used for generating an adjusting instruction according to the position deviation and the attitude deviation. According to the invention, precision limitation of traditional manual docking is broken through, intelligent management and control of the component docking process are realized, and docking precision and efficiency are improved.

The invention discloses a multi-intersection traffic signal control method fusing non-neighborhood information space-time perception. Traffic signal control of multiple intersections is optimized by adopting a reinforcement learning strategy, each intersection is used as a reinforcement learning agent, and space and time sequence characteristics of each intersection are extracted as state information. Non-neighborhood information spatial-temporal characteristics of multiple intersections are realized by using a non-neighborhood spatial characteristic enhancement graph transformer and a two-way multi-head cross attention mechanism to perform action decision, optimal control of a multi-intersection traffic network is realized, and spatial-temporal association of traffic flow and global road network situation awareness are realized.

The invention provides a construction progress intelligent dynamic identification and modeling method based on laser radar fused binocular and depth camera data. The method comprises the following steps: data acquisition: acquiring a construction site two-dimensional RBG image and three-dimensional point cloud data containing construction progress information through a multi-mode construction progress data acquisition system; performing data processing and analysis: performing fusion processing on the acquired two-dimensional RBG images and three-dimensional point cloud data by using a construction state recognition algorithm based on artificial intelligence, recognizing type information, spatial position information and time feature information of the components, performing construction progress quantification based on spatial volume difference to obtain construction progress information of each component, and performing data processing and analysis; forming a component point cloud data set; and performing three-dimensional progress dynamic reconstruction based on the building information model: dynamically reconstructing the progress information through the building information model BIM by using the component point cloud data set. According to the invention, dynamic and accurate evaluation and scientific guidance of the construction progress are realized through unmanned, intelligent and three-dimensional visualization means.

The present invention discloses a BIM model energy consumption simulation method and system coupled with a building performance database, which relates to the technical field of building information modeling and building green performance evaluation. The method establishes a building performance database template and inputs urban meteorological data, spatial operation data, and material physical parameter data; translates building site selection information, morphological information, spatial type information, and enclosure structure information to a digital design platform; calls building performance data from an external database according to the translated multi-level information; establishes a building energy model; and obtains the building's annual energy consumption data by simulation. Through the digital modeling platform and technology, the automatic translation of the multi-level information of the BIM model is realized, and the problem of insufficient support of the existing building information modeling technology for external related information is fully considered, and a dynamic transmission mechanism between the BIM model and external building performance big data is established.

Apparatus and methods are provided for CSI enhancement for multi-TRP coherent joint transmission. A user equipment (UE) receives signals from a plurality of transmission and reception points (TRPs). The UE determines, based on the signals, multiple TRP (multi-TRP) coherent joint transmission (CJT) channel state information (CSI) report information for a codebook based on a spatial basis selection matrix, a combination coefficient matrix, and a frequency basis selection matrix. The spatial basis selection matrix is layer common, polarization common, and TRP independent. The UE reports, to one or more of the plurality of TRPs, the multi-TRP CJT CSI report information.

The invention belongs to the technical field of distributed computing and data twinning, and provides a digital twinning method and system for the industrial Internet of Things, and the method comprises the steps: obtaining a multi-dimensional dynamic mapping model and a multi-modal time series data flow of a physical entity in the industrial Internet of Things in an information space; edge-cloud collaborative processing is carried out on the obtained multi-modal time series data flow, and fusion data used for driving the multi-dimensional dynamic mapping model is generated; performing hierarchical architecture deployment on the generated fusion data based on containerized micro-service; performing closed-loop control of industrial Internet of Things interruption on an edge side close to a physical entity by combining the obtained hierarchical architecture deployment and fusion data; and performing cloud data processing of the physical entity according to a closed-loop control result, and establishing a maintenance mechanism in which the edge and the cloud are synchronous and consistent, thereby completing digital twinning oriented to the industrial Internet of Things.

This application relates to the field of quality inspection technology, and discloses a method for human-machine interaction assembly quality inspection of a theodolite-type spaceborne laser coarse pointing mechanism based on HCPS. In the design stage, a design scheme is generated based on design requirements; in the manufacturing stage, the similarity of production tasks and the skill levels of manufacturing personnel are evaluated, tasks are matched with corresponding skills through strategy transfer, combined with AI technology and diverse interaction modes, feedback from manufacturing personnel is continuously collected and an adaptive strategy for human-machine interaction is implemented, and incremental learning based on events or historical data is autonomously carried out through a human-machine collaboration system; in the assembly stage, when the coarse pointing mechanism is assembled using equipment in physical space, based on the support of the information space, the quality of all components is checked before assembly; in the inspection stage, a test workstation is set up to offset the gravity influence of the mechanical structure, multiple parameters are measured, and the assembly quality is determined according to corresponding inspection principles. This application is based on human-machine collaborative CPM assembly quality inspection to ensure accuracy.

Systems and methods for providing a communication portal including an information space provided during a video conference are disclosed. In an example method, a computing system receives, from a first client device, a request to create a communication portal including one or more information spaces, each information space being filterable by category and including one or more information cards, each information card associated with one or more categories. The computing system receives a first indication to create a first information space associated with a first category. The computing system receives first information to be posted as a first information card within the first information space, in which the first information is related to the category. The computing system receives, from a second client device, a second indication to access the information space. The computing system provides the first information card to the second client device.