424 results about "Network computing" patented technology

The invention provides an underwater acoustic target recognition method based on autonomous task perception, and belongs to the field of underwater acoustic signal processing and artificial intelligence. S3, performing feature extraction on the time-frequency spectrogram by the task type extraction network to obtain a task embedding vector, calculating a similarity score, when the maximum similarity score is greater than a set threshold value, outputting a task representation vector, and entering S3, otherwise, inputting features output by the last Transform layer into a classifier; s3, selecting a router according to the task representation vector, selecting a trained expert network by the router, calculating a door control weight, processing universal acoustic features in parallel by the expert network, fusing output of the expert network, fusing deep features and fused features to obtain enhanced features, and enabling the enhanced features output by the last Transform layer to enter a classifier; the invention further provides a system. The problems that the task category autonomous recognition capability is insufficient and the correlation between tasks is ignored are solved.

The invention relates to the technical field of industrial mechanism models, in particular to a shock absorber performance optimization control method based on model fusion, which comprises the following steps: extracting a low-frequency disturbance state variable and inputting the variable-topology industrial mechanism model to generate a nominal reference state trajectory; utilizing a depth state observation network fused with energy passivity constraint to calculate a non-linear model mismatch compensation amount and an adaptive weighting parameter; performing dynamic fusion on the nominal reference state trajectory and the compensation amount based on the adaptive weighting parameter to generate a generalized state estimation value; and executing dynamic multi-objective optimization based on the generalized state estimation value, and generating mixed mode control input acting on an execution end. According to the invention, through adaptive fusion of a mechanism model and a data driving method, physical consistency, calculation real-time performance and robustness of a control process are considered.

The invention relates to the technical field of equipment energy consumption monitoring and fault diagnosis, and discloses a large-scale equipment energy consumption anomaly detection method and a storage medium. The method comprises the following steps: synchronously acquiring equipment energy consumption, state and environment data, fusing to generate multi-dimensional initial features, and performing multi-scale transformation. And performing preliminary detection on each scale feature subset to generate abnormal confidence and mode description. And screening the feature subsets according to the confidence coefficient, and dynamically selecting a corresponding detection strategy according to the mode description. And performing deep feature extraction on the feature subset based on the selected strategy, inputting the obtained high-dimensional feature vector into a corresponding anomaly evaluation network, calculating a standard state matching degree, comparing with a dynamic threshold to generate a fine-grained judgment result, and finally finishing positioning and attribution analysis in combination with anomaly description to form a detection report. According to the method, efficient and accurate large-scale equipment energy consumption anomaly detection is realized.

The invention provides a multi-modal semantic fusion image and text relevance dynamic analysis method and system, and belongs to the technical field of multi-modal data processing. The method comprises the steps of image and text data preprocessing, feature extraction, fusion semantic vector generation through a bidirectional cross attention mechanism, dynamic relevance score calculation through a time sequence attention long-short-term memory network and combined loss function end-to-end training. According to the method, accurate alignment of image and text features can be realized through a bidirectional cross attention mechanism, the cross-modal matching accuracy is improved, the dynamic correlation analysis capability is enhanced by using time sequence modeling, and the problems of insufficient feature coding suitability and limited time sequence modeling capability in the prior art are solved.

The invention discloses a flood peak enhanced physical base flow and residual error correction collaborative runoff prediction method, which belongs to the field of hydrological prediction, and comprises the following steps of: dividing a training set and a verification set according to a proportion, performing oversampling processing on flood peak samples, and constructing a time sequence window; a Xinanjiang model is discretized and expressed by adopting an ordinary differential equation, rainfall and potential evaporation data are input, and intermediate variables are obtained. A physical base flow and residual error correction dual-channel module is constructed, and physical base flow and residual error correction is calculated through two full-connection networks. And calculating a final runoff predicted value by adopting a residual connection structure, taking basic NSE loss as a core, superposing a flood peak sample error weighted item, strengthening flood peak fitting precision, and updating physical parameters and neural network weight through a back propagation algorithm. And verifying the model, and respectively calculating prediction indexes of the training set and the verification set. According to the method, fusion of a traditional hydrological model and a deep learning method is realized, the physical interpretation of the model is enhanced, and the basin runoff prediction precision is improved.

The invention discloses a port container automatic scheduling method based on multi-agent reinforcement learning, and the method comprises the steps: S1, building a corresponding relation between equipment and agents, and constructing a task set; s2, collecting operation state data, and constructing global and local state vectors; s3, generating a scheduling constraint vector, and cutting actions according to the resource, storage yard and path state to form a feasible action set; s4, on the basis of an improved QPLEX algorithm, constructing an individual value network containing a dump structure, and calculating an individual action value; s5, constructing a joint action value hybrid network, and mixing individual values according to the global state vector to form joint action values; s6, constructing a training sample, differentiating and aggregating instant and delayed return, and updating network parameters; and S7, during online scheduling, selecting an optimal action combination according to the combined action value, and generating and issuing a scheduling instruction. According to the invention, automatic collaborative scheduling of port container operation is realized.

The invention discloses a task processing method and device, in-network computing equipment, a medium and a product, relates to the technical field of computers, can provide a task processing method based on the in-network computing equipment, and improves the task processing efficiency. The method comprises the following steps: acquiring a task model and task data of a to-be-processed task from a central processing device by an in-network computing device, storing the task model and the task data of the to-be-processed task to the in-network computing device, performing block processing on the task data to obtain a plurality of task block data, and performing block processing on the task block data based on a plurality of processor cores of the in-network computing device. And performing parallel processing on the multiple pieces of task block data in the task model to obtain a task processing result.

The invention provides a multi-agent reinforcement learning regional energy collaborative scheduling method and system, and belongs to the field of regional energy system scheduling. Coupling degrees and a coupling degree matrix between agents are constructed; inputting the observation vector into a strategy network to obtain a decision action; individual basic rewards and system economic rewards are calculated, and constraint reference rewards are constructed; individual differentiation basic rewards are calculated, and rewards are distributed; inputting the decision action into a physical quantity prediction network, calculating a physical consistency reward, obtaining a final reward and a global reward, and calculating a target return; splicing observation vectors and decision actions of all agents, splicing global joint observation vectors and joint action vectors, inputting the spliced vectors into a value network, and training; inputting the local state set into the trained strategy network, outputting a scheduling instruction, inputting the scheduling instruction into the trained value network, and outputting an evaluation result; the problems of depiction rigidness of an intelligent agent coupling relation, lack of a cooperative benefit distribution mechanism and insufficient decision physical consistency are solved.

The invention relates to the technical field of industrial clusters, and discloses an industrial cluster industrial chain construction method and device, equipment and a medium, and the method comprises the steps: obtaining supply associated data and property right associated data of an enterprise of an industry category to which a target industrial cluster belongs; constructing a directed contact network based on the supply association data; constructing an undirected contact network based on the property right associated data; calculating property right connection degrees among the nodes in the undirected connection network, and accumulating the property right connection degrees to paths with corresponding node relations in the directed connection network according to a preset weight to generate a multi-source data enhanced network; and traversing all node paths in the multi-source data enhanced network, calculating the accumulated connection strength of each node path, and determining the path with the highest accumulated connection strength as the industrial cluster core industrial chain. According to the method, the universality, professional interpretability and objective accuracy of the constructed industrial chain are improved.

The invention relates to the technical field of office automation and business process intelligent supervision driven by big data, and discloses an office operation supervision method and system based on big data. The method comprises the following steps: firstly, extracting original logs containing an initiator, a receiver and timestamps from an approval system, checking one by one, de-duplicating and sorting according to time; counting flow times based on the ordered logs, and constructing an approval flow times matrix and a directed network; calculating the initiating and receiving quantity of each node, and screening actual participating nodes; calculating a flow imbalance index for each node, adaptively setting a threshold value according to a whole network average value and a standard deviation, and identifying abnormal nodes as bottleneck candidates; an adjacent matrix is constructed for the candidate nodes, connected clustering is carried out, and a structural bottleneck group is revealed; re-calculating the imbalance index according to an isochronous window in the whole monitoring period and counting the trend; and finally, combining the static unbalance degree and the dynamic change to generate a comprehensive score, and automatically outputting a structured monitoring report.

The invention relates to the technical field of information security, in particular to a network security event association detection method based on big data analysis. Comprising the following steps: data acquisition; feature extraction and fusion; correlation detection is carried out, wherein an improved Apriori-Bayesian fusion algorithm is adopted, and discretization processing is carried out on the event feature vectors; mining a frequent item set by using an improved Apriori algorithm; and risk assessment and result output. According to the method, an improved Apriori-Bayesian fusion algorithm is adopted, discretization processing is carried out on event feature vectors according to types, meanwhile, a security event weight factor is introduced to calculate the item set weighted support degree, and a minimum support degree threshold value is dynamically adjusted to mine a frequent item set; the association confidence coefficient is calculated in combination with the Bayesian network, and the confidence coefficient is corrected through the space-time association coefficient, so that the association relationship between the network security events can be scientifically judged, the problems of limited association judgment accuracy and lack of quantitative correction in the traditional technology are solved, and the association false alarm and missing report probability is reduced.

The invention relates to a coal machine spare part similarity identification and coding unification system and method and a storage medium. The system comprises a data acquisition module for extracting coal machine spare part data from a plurality of systems; the cleaning and preprocessing module is used for carrying out unified formatting processing on the spare part data; the feature extraction and matching engine is used for performing multi-dimensional feature extraction on the preprocessed data, including text similarity features, semantic features, structural features and attribute features; the similarity calculation and clustering module is used for calculating the comprehensive similarity of spare part data through a deep fusion network according to the multi-dimensional features, and performing automatic clustering according to the similarity; and the master data generation and system docking module is used for generating unified spare part master data for each cluster and synchronizing the spare part master data to a plurality of systems. According to the method, a large model and deep learning are utilized, multi-dimensional features are extracted and fused for multi-source spare part data, spare part similarity matching is carried out, dependence on a spare part coding format is broken through, and the accuracy and robustness of cross-system matching are improved.

The invention discloses a water conservancy project ecological construction method and system based on artificial intelligence, and relates to the technical field of ecological system simulation, and the method comprises the steps: collecting historical ecological event data of each region of a water conservancy project, analyzing the causal association between events based on an artificial intelligence model, and calculating and dynamically updating the association degree index between any two regions; constructing a regional association network based on the association degree indexes, calculating core scores of nodes in the network to identify a core region, and dynamically improving the monitoring sampling frequency of ecological data for the core region; and when an abnormal state is monitored in a single region, predicting the abnormal risk probability of other regions, and generating and recommending an optimal coping plan. Compared with the prior art, the method has the advantages that by constructing a dynamic regional correlation analysis system, the mutual influence between ecological regions is quantified, the monitoring resources are distributed uniformly to be intelligently and optimally configured based on the network core degree, and the monitoring efficiency and the capturing capacity of key information are remarkably improved.

Disclosed is a neural network computing device. The neural network computing device includes a neural network accelerator including an analog MAC, a controller controlling the neural network accelerator in one of a first mode and a second mode, and a calibrator that calibrating a gain and a DC offset of the analog MAC. The calibrator includes a memory storing weight data, calibration weight data, and calibration input data, a gain and offset calculator reading the calibration weight data and the calibration input data from the memory, inputting the calibration weight data and the calibration input data to the analog MAC, receiving calibration output data from the analog MAC, and calculating the gain and the DC offset of the analog MAC, and an on-device quantizer reading the weight data, receiving the gain and the DC offset, generating quantized weight data, based on the gain and the DC offset.

The invention discloses a neural network calculation method and device for gene expression regulation and control analysis, and relates to the technical field of bioinformatics, and the method comprises the steps: obtaining first feature data and second feature data; constructing an input feature comprising a plurality of regulation and control hierarchies; and inputting the input features of the plurality of regulation levels and the second feature data into the target neural network model, and outputting a prediction result of the gene expression state. According to the neural network calculation method provided by the invention, chromatin accessibility and three-dimensional space interaction data are deeply fused through a dynamic routing module, so that the problem of'black box 'which is inaccurate in prediction and difficult to explain in a traditional deep learning model is solved in a mode of explicitly simulating a real biological regulation mechanism; and a key gene regulatory pathway can be accurately identified.

The invention discloses an aircraft pulsation assembly line balance optimization method and device based on a graph neural network and deep reinforcement learning and a medium, and relates to the field of intelligent manufacturing, and the method comprises the following steps: S1, obtaining the initial information of an assembly line, constructing a disjunction graph model, and building a Markov decision process; s2, performing feature embedding extraction on the disjunction graph model by using a graph neural network to obtain node embedding and global embedding; s3, the nodes and the global embedding are input into an Actor strategy network, probability distribution is calculated to select a task to be assembled, and parameters of the graph nerve and the strategy network are dynamically adjusted; and S4, judging whether the current assembly line balance scheme meets a preset condition, if so, outputting the scheme, and otherwise, returning to the step S2 to continue training. The scheduling deviation caused by incomplete model expression is avoided, the modeling mode provides a solid foundation for real-time decision making, and the response speed of the assembly line to the dynamic production environment is increased.

The present disclosure relates to the technical field of optical computing, and particularly relates to an intelligent optical computing chip cluster architecture and system. The architecture comprises: a plurality of optical computing chips, and optical path channels are established between the plurality of optical computing chips by using spatial phase adjustment to modulate light beams; and high-speed optical communication and optical diffraction network computing are performed between the plurality of optical computing chips by performing time-domain intensity modulation on the optical path channels. The present disclosure adopting the above scheme can realize the interconnection computing of optical computing chips in the spatial dimension, and does not rely on electricity for information transmission.

The invention discloses a multi-modal neural network calculation method for predicting the residual life of an ion tube, and the method comprises the following steps: analyzing the operation state characteristic data of the ion tube, and extracting the modal characteristics of the data; constructing and training a multi-modal neural network; and predicting the residual life of the ion tube by adopting a multi-mode neural network. The designed method achieves prediction calculation of the residual life of the ion tube, and has the advantages of being matched with the actual state and accurate in calculation.

This application describes a network-on-chip system that could be used in a hardware accelerator for accelerating neural network computations. An example NoC system may include a plurality of cores, and a plurality of data links connecting adjacent cores of the plurality of cores for transmitting data. The example NoC may further include a global synchronization switch connected to each of the plurality of cores. The global synchronization switch is configured to dynamically connect any pair of cores in the plurality of cores for transmitting control signals between the pair of cores.

The invention provides a building facility real-time monitoring and early warning method and system based on the Internet of Things, and relates to the technical field of building monitoring and early warning, and the method comprises the steps: obtaining real-time monitoring data through multiple sensors, and carrying out the sliding analysis and wavelet decomposition of the data, and obtaining a structural feature vector; calculating a spatial correlation weight by using a self-attention network, and clustering the spatial correlation weight; executing multi-scale decomposition to calculate an abnormal score; calculating a structure deviation degree by combining a variational encoder; and generating early warning information when the deviation degree of a plurality of continuous monitoring periods exceeds a threshold value. According to the invention, accurate positioning and early warning of building damage are realized, and the monitoring efficiency and safety are improved.

The invention relates to the technical field of government affair services, in particular to an NLP entity recognition method based on semantic extension, which comprises the following steps: collecting multi-source time sequence data and generating an entrepreneurship ecological data graph through a dynamic time warping algorithm; dynamic evolution of a term library is realized by using a generative adversarial network, and a term semantic association graph is constructed in combination with a graph convolutional network; a double-path attention mechanism is deployed in a Transform encoder, and global semantics and local attention guided by terms are fused through a gating loop unit; modeling fuzzy query analysis into a Markov decision process, and optimizing a semantic extension strategy by adopting a reinforcement learning agent; multi-granularity features are extracted through a feature pyramid network, and model parameters are optimized in combination with a bidirectional long-short-term memory network and an online learning feedback loop; and finally, calculating entity service association strength by using a graph attention network, and generating an interpretable report based on a Shapley value attribution algorithm. A closed-loop learning system is formed, and the business starting policy matching accuracy is improved.

The invention relates to the technical field of wireless communication, in particular to a wireless communication system signal sending and receiving method based on an AI model. The system comprises a delay coordinate embedding unit, a neural flow tracking unit, a continuous evolution unit, a Riemann decoding unit and a closed-loop control unit. The system maps the pilot frequency to a high-dimensional phase space by using a Takens principle, and reconstructs a channel dynamic topology; the core of the method is that a manifold derivative network is utilized to calculate a flow field vector, continuous time deduction is carried out through a Shenchang differential equation, evolution deviation is eliminated in combination with a geometric projection correction term, and thus the future channel state is accurately predicted. Calculating a CSI matrix based on the prediction result to optimize beam transmission; according to the method, the limitation of single-variable observation is broken through, and reconstruction and perception from discrete data to a full view of a high-dimensional nonlinear system are realized.

The invention belongs to the technical field of computers, and particularly relates to a personalized recommendation method and system based on a dynamic heterogeneous graph and reinforcement learning. The method comprises the following steps: firstly, constructing a global heterogeneous information graph of multiple types of nodes offline, and learning static embedding of the nodes by using a graph neural network; secondly, dynamically constructing a session history into a session graph in a real-time interaction process of the user, and aggregating by adopting a graph convolutional network to generate a dynamic state vector of the user; inputting the dynamic state vector into an actor and commentator reinforcement learning framework; and finally, using a dominant function calculated by the commentator network as a stable learning signal, and performing end-to-end joint training on the whole model to optimize long-term cumulative return. According to the method, by introducing the session graph volume accumulation device, the accuracy of dynamic state representation is remarkably improved; and an actor commentator framework is adopted, so that the problem of high variance of a traditional strategy gradient method is effectively solved, and the training stability and efficiency are improved.

Techniques are disclosed relating to exchange of images in networked computing applications. In particular, the disclosure relates to exchange of gain curves that are used to represent imaging and / or video in such applications. A gain curve may define a mathematical transformation that relates values from a source image domain to a destination image domain. The image and its associated gain curve(s) may be published to destination devices for consumption. When a destination device consumes the image, the destination device may apply a transform to source image content according to the gain curve(s) published with the image. For example, the destination device may apply a gain curve to an associated image directly, or it may derive another transform from the gain curve and additional information known to the destination device.

The invention relates to an incremental learning CNN-Attention-based broadband impedance on-line identification method for a network construction type wind power plant, and belongs to the technical field of new energy grid connection, and the method comprises the following steps: S1, building a parallel collection type offshore wind power plant multi-machine dynamic equivalent model; s2, collecting and preprocessing parameters based on a multi-machine equivalent model of the network-building type wind power plant; s3, a CNN-Attention model is constructed and trained; s4, constructing an equivalent impedance model of the current collection submarine cable line between the wind turbine generators; s5, on the basis of the actual topology of the network construction type wind power plant station, connecting the broadband impedance models of the wind turbine generator and the current collection circuit into a frequency domain impedance network, and calculating the impedance characteristics of the current collection port of the multi-machine equivalent wind power plant; s6, a CNN-Attention model containing an incremental learning updating mechanism is constructed; and S7, performing broadband impedance on-line identification on the network construction type wind power plant.

Techniques and architecture are described for a method that includes an in network compute (INC) manager receiving from switches of a fat tree configured network, arithmetic logic unit (ALU) capacity of the switches. Based at least in part on the ALU capacity of the switches and bandwidth, the INC manager determines one or more switches within each tier that are capable of supporting the processing units and based at least in part on the determining, the INC manager selects a first switch as a root, wherein the first switch is included within a tier of switches having intermediate tiers of switches located between the tier and the plurality of processing units within the fat tree configured network. The INC manager creates one or more paths of switches within each of the intermediate tiers from the root to the plurality of processing units to provide a constrained disjoint spanning tree of switches.

The invention discloses a safety reinforcement learning automatic driving decision-making method based on a behavior correction mechanism, and belongs to the technical field of automatic driving decision-making. Comprising the following steps: S1, collecting historical driving data to construct a mixed data set, and training a safety evaluation network according to the mixed data set; s2, calculating a safety score based on the trained safety evaluation network, and executing an action or triggering a correction mechanism according to the safety score and a dynamic safety threshold; s3, storing empirical data according to a security level by adopting a hierarchical empirical playback buffer area, and updating a strategy in combination with a constraint optimization and regularization method; s4, safety parameters are adjusted based on the real-time environment state, safety indexes are counted through a sliding window, and a basic safety threshold value is adjusted. By adopting the safety reinforcement learning automatic driving decision-making method based on the behavior correction mechanism, the safety violation frequency in the training process is reduced, the task completion efficiency is improved, and the method is suitable for automatic driving decision-making control in a complex traffic scene.

The invention particularly relates to a target classification and uncertainty evaluation method based on semantic association evidence fusion, and the method comprises the following steps: 1, constructing and training a target fusion classification neural network, and calculating a Dirichlet distribution concentration parameter of each modal input data representing an evidence quantity; step 2, associating the generated Dirichlet concentration parameter with the evidence quantity, and calculating single-mode uncertainty; step 3, constructing a semantic association matrix and performing discount correction, exploring potential association and confusion relationships among different categories, and completing multi-source evidence fusion through a Dempster combination rule; step 4, integrating the global uncertainty of the fused evidence and the local uncertainty of the single mode to obtain final uncertainty evaluation; according to the method, the modal information can be effectively fused to obtain high-precision fusion classification, the uncertainty of fusion classification can be quantitatively evaluated, a basis is provided for improving the safety and interpretability of intelligent classification decision, and the method is suitable for the multi-source sensor decision fusion field of automatic driving, medical diagnosis and the like.

The present disclosure discloses a method for detecting anomalies in acroengine variable working conditions based on a multi-band wavelet network, including the steps of: collecting monitoring signals of an acroengine in a healthy state under a plurality of working conditions and corresponding working condition information; constructing and training a multi-band wavelet network using the signals and working conditions; calculating a state monitoring index of the signal using the trained network; determining a monitoring threshold according to the state monitoring index; and determining whether or not the engine is abnormal by comparing the state monitoring indexes of the monitoring signals of the aeroengine under the plurality of working conditions to the monitoring threshold. The present disclosure fully considers the advantages and disadvantages of the signal processing technique with the deep learning network for fault diagnosis.