182 results about "Interaction network" patented technology

The invention relates to the technical field of medical monitoring and early warning, and discloses a postoperative care risk early warning system for neurosurgical patients. According to the system, real-time monitoring flows of electroencephalogram, intracranial pressure and body surface myoelectricity of a patient are continuously received, and a dynamic interaction network of neuroelectrophysiology and intracranial mechanical states is constructed so as to represent the coupling relation among electroencephalogram rhythm, pressure conduction and postoperative skull windowing area tissue compliance. The system analyzes three types of core parameters of cerebral cortex excitability, cerebrospinal fluid circulation load and cranial cavity compensation space from the network, further calculates a brain tissue perfusion risk and a neural structure compression risk, and generates a corresponding nursing early warning signal and a physiological regulation path. According to the scheme, deep coupling and mechanism risk early warning of multiple physiological signals are achieved, and the postoperative complication risk can be recognized earlier and more accurately.

The invention discloses a video crowd counting method based on a cascaded cross-domain feature interaction network. The method comprises the following steps: carrying out data enhancement processing of random cutting and horizontal flipping on a current frame and front and back frames of the current frame; and constructing a cross-domain feature interaction network composed of a spatial domain branch and a frequency domain branch. The frequency domain branch extracts frequency domain feature output of different stages through a high and low frequency signal aggregation module and a feature encoder based on adjacent frames; the spatial domain branch is based on a single-frame image, and static spatial semantic features are extracted through a feature encoder. Cascade fusion is carried out on the double-branch features on multiple scales, two-way channel cross attention is utilized to reconstruct time sequence correlation frequency domain features of a current frame, and fusion and reconstruction of the two domain features are achieved through a cross-domain feature mutual modulation module. And after the reconstructed double-branch features are processed by the fusion network, outputting a crowd density map of the current frame by a density regression head. And after training is completed, storing the optimal model for video crowd counting. According to the invention, through cross-domain feature cascade and bidirectional time sequence modeling, the accuracy and robustness of crowd counting in a video scene are effectively improved.

The invention discloses a fleet formation flight implementation method and device, equipment, a medium and a product, and belongs to the technical field of aeronautical communication. The method comprises the following steps: establishing an air-ground interconnection communication link to realize real-time data interaction between a ground control unit and a plurality of flying civil aircrafts and between the plurality of civil aircrafts; based on an air-ground interconnection communication link, an airborne ad hoc network with a leading aircraft as a core is constructed, and an information interaction network in the formation is formed. Acquiring formation flight environment data; determining a safe area of formation flight based on the formation flight environment data, and calculating the maximum number of aircrafts in the formation and the relative positions of the aircrafts in the formation; and issuing the maximum number, the relative position and the flight decision information to each airplane in the formation through the airborne ad hoc network so as to execute formation flight. According to the embodiment of the invention, the real-time performance and accuracy of communication between the fleet can be effectively improved, and the airspace operation efficiency and flight safety are improved through the formation flight of the fleet.

The invention discloses a multi-modal data joint embedding method based on a hierarchical progressive arithmetic interaction network, and the method comprises the following steps: obtaining text data and image data from the same semantic entity, and extracting a text local feature, a text global feature, an image local feature and an image global feature; inputting the text local feature and the image local feature into an atomic layer, and processing based on a Cartesian product to generate a first-order interaction feature; inputting the first-order interaction features into a combination layer, and carrying out nonlinear transformation processing to generate enhanced nonlinear interaction features; and inputting the nonlinear interaction features into the aggregation layer, and generating a multi-modal joint embedding vector in combination with the text global features and the image global features. The method effectively solves the problems of modal isomerism, single interaction level, lack of dynamic adaptability and the like in the prior art.

The invention relates to the technical field of computer mode recognition, and discloses a multi-source heterogeneous data fusion processing and key feature extraction method and system, and the method comprises the steps: achieving the adaptive caching and granularity normalization of streaming data through a dynamic buffering queue and a time alignment window; generating a structured vector of a unified space-time reference by using a structured analysis module; a high-dimensional fusion feature tensor is constructed through two-stage convolutional coding and a cross-source attention interaction network; and a key feature channel is screened based on gradient sensitivity through a differentiable channel pruning module. The system comprises a multi-source data access unit, a dynamic buffer management unit, a time alignment unit, a synchronous resampling unit, a structured analysis unit, a primary fusion coding unit, a cross-source attention interaction unit, a time sequence dependence modeling unit, a feature importance evaluation unit, a key feature screening unit and the like. According to the method, efficient, accurate and low-overhead multi-source heterogeneous data real-time fusion and task-oriented key feature extraction can be realized.

The invention provides a multi-modal cerebral arterial thrombosis medical image segmentation method, device and equipment, and the method comprises the steps: extracting the independent features of different modal medical images through combining a ViT encoder branch and a CNN encoder branch which are finely adjusted by a hybrid expert as a multi-modal image double-branch coding network; further integrating complementary information of different modes by using a mode missing adaptive fusion network, and performing global-local information interaction between CNN features and ViT features by using an encoder branch interaction network, so that specific features and cross-mode invariant features of different available modes can be decoupled under the condition of mode missing; and meanwhile, the advantages of different types of features are fully utilized, and the value information of the multi-modal image features is deeply mined, so that accurate multi-modal cerebral arterial thrombosis medical image segmentation and imaging are realized, and the method is high in reliability, good in accuracy and good in practicability.

The invention discloses an industrial intelligent safety risk early warning and management and control platform, and relates to the technical field of industrial safety monitoring and management and control. The industrial intelligent security risk early warning and management and control platform comprises a full-scene perception layer, a data management layer, an intelligent risk identification layer, a hierarchical management and control layer and a global operation and maintenance layer, and each layer and an edge computing node construct a distributed data interaction network through an industrial Ethernet; the scene sensing layer is used for realizing all-region and all-time security data acquisition of industrial production and comprises a multi-mode sensing terminal array, an equipment protocol adaptation module, a personnel state monitoring unit and an acquisition scheduling module; according to the industrial intelligent safety risk early warning and management and control platform, through a multi-mode terminal array and a dynamic acquisition scheduling mechanism of a full-scene sensing layer, full-area and full-time-period automatic monitoring of industrial production is realized, dependence on manual inspection is thoroughly eliminated, the problems of low manual inspection efficiency and missing inspection of hidden risk points are effectively solved, and the industrial production safety risk early warning and management and control platform is suitable for popularization and application. And the monitoring coverage rate is improved to 100%.

The invention discloses a transport capacity resource integrated intelligent scheduling method and system oriented to multiple service scenes, and relates to the field of intelligent scheduling, and the method comprises the steps: constructing a hierarchical collaborative decision-making architecture comprising a macroscopic strategic layer agent and a microscopic tactical layer optimizer; and inputting the structured feature vector into a hierarchical collaborative decision-making architecture, dynamically distributing proper solution algorithms and parameters for a microscopic tactical layer optimizer according to a real-time scheduling situation by utilizing an online element learning optimizer, and outputting a pre-scheduling scheme. According to the invention, through integrated integration of multi-service scene data, comprehensive utilization of static basic information, real-time transport capacity data and prediction environment data is realized, and the data support capability of scheduling decision is improved. The feature interaction network of the multi-layer perceptron structure can accurately extract core features and provide effective input for scheduling decisions. The hierarchical collaborative decision-making architecture is combined with an online element learning optimizer, so that a solution algorithm and parameters can be dynamically matched, and a better pre-scheduling scheme can be output.

The invention provides a blue-green algae image recognition method and system based on hierarchical self-adaption and domain driving, and the method comprises the steps: enhancing an image through employing an improved dark channel algorithm; constructing a blue-green algae biological attribute text database, and performing synonym replacement and sentence pattern recombination; multi-scale visual features are extracted through a hierarchical adaptive Swin Transform model, and key region characterization is enhanced in combination with dynamic spectrum attention; the text is input into a Bio-ALBERT model, and semantic embedding of field optimization is generated through term mask prediction and attribute relation pre-training; constructing a two-layer heterogeneous graph by using a graph attention interaction network GAIN, calculating a cross-modal association weight through a bidirectional graph attention mechanism, and outputting a cross-modal graph feature; multi-scale cross-modal association is modeled through a hierarchical graph attention fusion mechanism, and a comparison alignment loss optimization model is combined; and high-precision blue-green algae identification is realized. According to the method, through multi-scale perception, domain semantic adaptation and graph structure fusion, the accuracy of blue-green algae detection in a complex environment is improved.

The invention discloses a traffic scene training data generation method and device, electronic equipment and a medium, and relates to the technical field of intelligent traffic, and the method comprises the steps: obtaining multi-modal traffic data, building a dynamic semantic interaction network according to the multi-modal traffic data, and enabling the dynamic semantic interaction network to comprise a plurality of traffic entities, the dynamic attribute of each traffic entity and the space-time relationship between the traffic entities are determined; in the dynamic semantic interaction network, labeling the target event and a core node of the target event according to a preset event specific sub-graph; performing causal chain backtracking on the labeled target event and the core node of the target event to obtain structured causal chain data corresponding to the target event; and performing question and answer pair generation processing on the structured causal chain data corresponding to the target event to obtain an instruction fine tuning data set for training the traffic large model. Therefore, automatic and high-quality generation of the traffic scene training data is realized, and the logical reasoning ability and interpretability of the model are improved.

The invention relates to the technical field of network security, and discloses a server data interaction network security monitoring processing method and device, in a TLS / SSL handshake process between a server and a client, non-sensitive parameters in a key negotiation process are extracted to generate a key fingerprint, pre-risk assessment is carried out in combination with a preset risk knowledge base, and the key fingerprint is obtained. Obtaining an initial risk score; according to the method, non-sensitive parameters are extracted in a TLS handshake stage to generate key fingerprints, and dynamic association analysis is performed on the key fingerprints and subsequent ciphertext data transmission behaviors, so that a novel detection normal form which does not decrypt but insight into risks is constructed; when a novel encryption attack occurs, a time-consuming traffic decryption operation does not need to be carried out, and accurate threat identification can be realized only by analyzing abnormal association of handshake features and behavior patterns, so that the problems of performance bottleneck and privacy compliance of security detection in an encrypted traffic environment are fundamentally solved, huge calculation overhead caused by decryption is avoided, and the security detection efficiency is improved. The method is suitable for high-speed network environments.

The invention relates to a data processing technology, and provides a mass distribution network industrial resource collaboration method based on load interaction and response analysis, which comprises the following steps: firstly, constructing a real-time synchronous data acquisition and processing system, and fusing multi-source data to form a unified data base; secondly, realizing flexible dynamic identification through mechanism modeling and data driving, and introducing a flexible label to quantify adjustable power, duration and response rate; constructing an interactive network based on the flexibility label, depicting a space-time coupling effect and group response characteristics, and generating an equipment, station and region level scheduling scheme through a multi-target hierarchical optimization model in combination with market electricity price and renewable output prediction; and finally, comparing the plan with the actual response in the actual scene, and performing correction. According to the invention, the cooperative scheduling and fine management capability of the power distribution network on massive industrial loads can be improved, the flexible utilization level is improved, and the economical efficiency and robustness of the system are enhanced.

The invention provides a Mini LED module service life prediction method and device based on operation data driving and a storage medium, and relates to the technical field of display.The method comprises the steps that firstly, multi-source operation data and full-period historical service life associated data of a Mini LED module are collected, and then a target service life influence factor is used as a core node to predict the service life of the Mini LED module; the method comprises the following steps: constructing a life consumption interaction network for a flexible connection edge according to a dynamic interaction relationship between factors, inputting multi-source operation data into the network, carrying out cross-dimension dynamic interaction coupling on the multi-source operation data and full-period historical life associated data to generate a real-time historical interaction coupling result, and generating a self-adaptive life consumption rate according to a dynamic adaptive node response mode; and finally, generating a residual life prediction result according to the self-adaptive life consumption rate and the current comprehensive performance state data of the module, and reversely inputting the subsequent actual data into the network to dynamically adjust the interaction relationship, so that the life of the MiniLED module can be accurately predicted.

The invention discloses a sheep infectious disease number zero individual tracing method and system based on a group interaction network, particularly relates to the technical field of agricultural informatization, and is used for solving the problem of number zero individual misjudgment caused by neglecting group dynamic behaviors in an existing static network model. The method comprises the following steps: dividing a continuous time window to construct a time sequence interaction network by acquiring dynamic contact data and high-risk medium use data of a sheep flock; virtual edges are added to individuals with the use interval smaller than the pathogen survival time of the same high-risk medium; calculating a node behavior mode mutation degree based on a medium using interval variance, and analyzing propagation fluctuation in combination with an input / output propagation flow ratio standard deviation; identifying nodes of which behavior mutation exceeds a threshold value, propagation fluctuation continuously exceeds the threshold value and output propagation flow suddenly drops as an abnormal propagation source; and extracting a first window associated sub-network, and judging the node which has the highest behavior mutation degree and uses a high-risk medium in the sub-network as a zero individual, thereby realizing accurate tracing of the sheep infectious disease source and providing a basis for accurate prevention and control of a farm.

The invention provides a dynamic gesture recognition method based on lightweight multivariable space-time convolution, and the method comprises the steps: constructing a spatial feature extraction module based on a pseudo 3D gated attention fusion network, extracting multi-scale spatial features through the spatial feature extraction module, and injecting a guide heat map through a gated attention fusion module, enhancing features of the key region and inhibiting background interference to obtain a space refined feature sequence; and decomposing the spatial refined feature sequence into a plurality of sub-variables, parallelly capturing long-range and local time dependence by using a modern convolution module, and respectively carrying out relation modeling in the variables and between the variables through a decoupling feature interaction network to obtain a dynamic gesture recognition result. According to the method, a multivariate characteristic decomposition strategy is combined with modern convolution, the double-branch design advantage of the modern convolution is that long-range dependence and local details are taken into consideration, the limitation of traditional convolution in the aspect of capturing a long-range time dependence relationship is solved, and modeling is performed on a complex dynamic state.

The invention discloses a newborn inherited metabolic disease risk assessment method based on multi-modal collaborative learning, and the method comprises the following steps: collecting and preprocessing gene, metabolism and phenotype data, and constructing normalized multi-modal input features; constructing a cross-modal interaction network, and extracting fusion features from the preprocessed data to enhance the expression ability of pathological information; constructing a metabolic pathway diagram, and modeling a structure and function relationship between metabolites by using a diagram convolutional network; and performing feature alignment on the multi-modal features and a metabolic pathway diagram representation vector, and realizing efficient multi-modal fusion reasoning based on a confidence-guided incremental attention mechanism to complete disease data processing. Through multi-modal data fusion and graph convolutional network modeling, the pathological information expression ability is enhanced, efficient fusion reasoning is realized, and the accuracy and efficiency of neonatal disease data processing are improved.

This invention provides a method and system for assessing the impact of construction disturbance on biodiversity in nature reserves, relating to the field of computer technology. The method includes: acquiring species distribution data, construction disturbance data, and environmental data within the nature reserve; constructing an ecological association network based on the species distribution data to obtain a species interaction network; extracting network structure features from the species interaction network to obtain a set of key species; modeling the disturbance propagation dynamics based on the set of key species and the construction disturbance data to obtain an impact propagation path map; predicting biodiversity responses based on the impact propagation path map and environmental data to obtain trends in changes in biological community composition; and conducting a comprehensive impact assessment based on these trends, outputting an impact level index by comparing the degree of deviation from historical baseline states. This invention effectively improves the overall accuracy and predictive precision of the assessment of the impact of construction disturbance on biodiversity in nature reserves.

The application relates to the technical field of network security, and discloses a server data interaction network security monitoring processing method and device, wherein in a TLS / SSL handshake process between a server and a client, a non-sensitive parameter in a key negotiation process is extracted to generate a key fingerprint, and an initial risk score is obtained by combining preset risk knowledge bases to perform pre-risk evaluation; the application extracts the non-sensitive parameter to generate the key fingerprint in the TLS handshake stage, and dynamically analyzes abnormal association of handshake features and behavior modes in combination with subsequent ciphertext data transmission behaviors, so that a new detection paradigm for risk insight without decryption is constructed; when a new encryption attack occurs, time-consuming traffic decryption operation is not needed, and accurate threat identification can be realized only by analyzing abnormal association of handshake features and behavior modes, so that the performance bottleneck and privacy compliance problem of security detection under an encrypted traffic environment are fundamentally solved, huge calculation overhead caused by decryption is avoided, and the application is suitable for high-speed network environments.

The invention relates to the technical field of social network application, in particular to a social media viewpoint evolution simulation method and device based on coupling dynamics, and the method comprises the steps: extracting event topic interaction network data based on social network public information, carrying out the statistics of user historical interaction behaviors, and determining an information transmission network and an internal association network according to the data, the two forms a multi-layer association network; determining a transmission state of forwarding and a text viewpoint baseline for calculating a transmission state and a viewpoint value of a user in the social network data set; and inputting the propagation state and the viewpoint value into a coupling dynamics simulation model, carrying out iterative calculation until the viewpoint value converges so as to obtain a final propagation state and a final viewpoint value of the user, and generating forwarding situation distribution and user viewpoint distribution. Therefore, the problems that errors are generated, public opinion monitoring and early warning and public opinion guide strategy construction are affected and the like due to the fact that a viewpoint evolution model adopts a single-layer propagation structure and coupling modeling is not carried out on propagation and an internal correlation structure in the related technology are solved.

The invention relates to the technical field of historical building structure health monitoring, and discloses a historical building safety performance evaluation method and system. The method comprises the steps of obtaining a structure response signal through monitoring of a multi-position sensor, and generating a fusion feature map through a multi-scale feature sensing network. And performing three-dimensional reconstruction based on the atlas, adaptively dividing difference precision entity units according to feature density, and constructing the high-fidelity digital twinborn body. And driving the twin to perform time-history evolution under simulation of long-term environment excitation, and extracting damage evolution behavior fingerprints. And performing multi-round iterative comparison and reverse traceability analysis on the behavior fingerprint and a safety baseline containing material time-varying degradation and component interaction influence, identifying a failure mode and an evolution path, calculating a residual bearing capacity margin of a key component, and generating a structural life map in combination with a component interaction network. According to the method, high-fidelity modeling based on monitoring data and reverse intelligent diagnosis of dynamic damage evolution are realized.

The present application relates to the field of OTN resource matching, and more particularly to an OTN resource intelligent matching method based on artificial intelligence, comprising: inputting the preprocessed service demand, available wavelength resource, real-time bandwidth utilization and support OTU level into a feature extraction model based on a residual multi-layer perception architecture to generate a prediction base feature; inputting the prediction base feature into a feature interaction model based on a compressed interaction network architecture to generate a service comprehensive interaction feature; and inputting the service comprehensive interaction feature into an OTN resource matching model based on an expert gate network architecture to generate a recommended OTU level. Through the hierarchical intelligent model processing of the residual multi-layer perception architecture, the compressed interaction network architecture and the expert gate network architecture, the present application realizes the significant improvement in the accuracy, efficiency, adaptability and network operation value of OTN resource matching.

The invention discloses a method for constructing, updating and retrieving an action memory library, which belongs to the technical field of computer vision and comprises the following steps of: constructing a training data source with time sequence diversity; initializing an action memory library containing a plurality of learnable prototype matrixes and a double-flow interaction network; memory bank evolution is executed, an action prototype is retrieved by utilizing a query stream, a current memory state is dynamically generated by combining a memory state updating gate mechanism with a historical state, and dynamic memory is injected into a feature space by utilizing memory driving graph convolution; synchronously updating parameters based on multi-target loss, and driving a memory bank to evolve into optimal structured prior; and finally, performing structured reasoning on the to-be-detected sequence by using the optimal memory bank. According to the method, structured priori is constructed by mining a spatio-temporal topology mode of a human body action sequence, and hierarchical memory evolution and double-flow depth interaction are combined, so that the problem of depth ambiguity in a monocular vision task is effectively solved, geometric structure distortion is corrected, and the accuracy of action posture estimation is remarkably improved.

The invention relates to the technical field of seabed engineering, and discloses a submarine cable construction simulation method under complex seabed geological conditions. The method comprises the following steps: slicing and recombining multi-source investigation time sequence data, and constructing a geologic feature space based on a multi-dimensional time sequence window and time-space synchronization. Through feature fusion and dimension mapping, three-dimensional geological semantic body units with consistent time and space are generated, and attributes of the three-dimensional geological semantic body units and construction process parameters are associated and coded. And a recursive segmentation algorithm is adopted, and layered and nested dynamic construction decision units are automatically divided according to an attribute parameter mutation threshold. On the basis, a simulation agent is initialized for each decision-making unit, and dynamic re-evaluation of attribute parameters is driven through an agent interaction network. According to the method, the fidelity of the simulation model to the space-time evolution characteristics of the complex geological conditions is improved, the construction decision can adaptively respond to the local mutation of the geological parameters, and the simulation accuracy and the engineering practicability are enhanced.

The invention belongs to the technical field of pollution control, and relates to a soil pollution treatment method and system utilizing microbial remediation, and the method comprises the following steps: obtaining target remediation functional microbial agent genome and to-be-remedied site native microbiome metagenome data, and constructing a biological information basic data set; processing the data, and constructing an interaction network model containing microbial inoculum and native species nodes based on metabolic complementarity and ecological niche overlapping degree simulation calculation; analyzing the network topology structure to screen a key native co-generation node set, and generating a growth promotion demand map; determining a targeted metabolism regulation factor with targeted selectivity based on atlas reverse matching, and generating a targeted signal instruction; executing the instruction and putting a microbial agent, activating a synergistic node in situ and coupling with the microbial agent to construct a degradation function network; the method solves the problem that the colonization efficiency of the exogenous functional microbial inoculum is not high due to lack of accurate regulation and control on the native microbial community.

The invention relates to a soil organic matter prediction method based on multi-feature fusion, and belongs to the technical field of soil organic matter prediction. The method comprises the following steps: preprocessing acquired soil data to obtain spectral features, and performing time domain reconstruction of frequency domain signals to obtain time domain data; obtaining an optimal delay time and an optimal embedding dimension based on the time domain data, and performing phase space reconstruction to obtain a phase space trajectory; chaos features are extracted based on the phase space trajectory, and the extracted chaos features, the optimal delay time and the optimal embedding dimension serve as final chaos features; obtaining a vegetation index based on the spectral feature and taking the vegetation index as an index feature; and inputting the spectral features, the final chaotic features and the index features into a constructed double-flow low-rank interaction network model to obtain a soil organic matter prediction result. The objective of the invention is to solve the technical problem of low prediction precision caused by the fact that spectral features extracted in the prior art cannot comprehensively represent complex nonlinear characteristics of soil.

The invention belongs to the technical field of image processing, and particularly relates to a stereoscopic vision parallax prediction method based on an energy function and an attention fusion network. The objective of the invention is to solve the problem that a real-time algorithm is difficult to obtain high precision in stereoscopic vision parallax prediction. The method comprises the following specific steps: acquiring a data set: training a network model by adopting two public virtual and real stereo matching data sets; the method comprises the following steps of: constructing a network model: constructing an Entry Function and Attention Interference network, and optimizing cost body construction and cost aggregation by utilizing an attention mechanism; designing a minimum loss function; training a network model: inputting the acquired data set into the network model for training; finely adjusting model parameters; and storing model parameters. According to the method, an end-to-end supervised method is adopted, a UNet-like network architecture is integrally used, cost body construction and cost body aggregation are improved, and the efficiency of the model is improved by using an attention mechanism, so that the real-time stereoscopic vision parallax prediction method based on deep learning is realized.

The invention relates to an intelligent logistics POI recommendation method based on semantic knowledge distillation and interpretability. The method comprises the steps of obtaining data for training; according to a feature engineering module, encoding each type of data, inputting obtained user identification features, POI attribute features and user behavior features into a deep interaction network of a teacher model, and calculating a preliminary interaction score between a user and a POI based on node features of user nodes and POI nodes after updating of a graph neural network; a lightweight model is used as a basic framework of a student model, semantic knowledge of a teacher model is inherited through adaptive semantic knowledge distillation, test data is input into the student model inheriting the semantic knowledge, final interaction scores of a user and POIs are output, a plurality of corresponding POIs with the highest final interaction scores are used as recommendation results, an SHAP value is calculated, and a recommendation result is obtained. And the recommendation result is explained. According to the method, efficient recommendation, semantic maintenance and decision transparency can be realized at the same time.

The application discloses a large model system based on a computing acceleration chip, and relates to the field of large models. A plurality of computing acceleration units and a management server are used to form and implement an inter-chip and off-chip data interaction network; a hybrid display memory of the computing acceleration unit is matched with an SSD to form a multi-source storage mode; the computing chip is internally provided with a normalized on-chip network, an interconnection transmission system, a storage control system, and a plurality of computing acceleration cores; the normalized on-chip network can read model parameters of a target position based on the storage control system and send the model parameters into the computing acceleration cores for calculation and storage; the interconnection transmission system is used to interact with the management server and the remaining computing acceleration units, and read and store external model parameters and inter-chip model parameters. Through the collaborative design of the hybrid display memory architecture and the normalized on-chip network, in combination with a multi-level routing control and a dynamic configuration mechanism, the technical problems of insufficient display memory capacity, excessively high hardware cost and limited transmission bandwidth in the traditional scheme are effectively solved.

The application discloses a kind of prediction method of microorganism and drug relationship based on multi-association graph, through microorganism-drug association database, the association network of microorganism-drug is constructed, and further obtains interaction network Net1, Net2 and Net3;Establish the graph neural network model of introduction regularization, input graph neural network model with Net1, Net2, Net3 combine the multi-modal attribute graph of microorganism-drug to obtain embedding representation Z1, Z2 and Z3, and embedding representation Z1, Z2 and Z3 are input into graph neural network and trained, to obtain trained graph neural network;Finally, the microorganism-drug association effect in microorganism-drug data set is predicted by trained graph neural network model.The application constructs the node feature of biological and drug that can be explained, and considers the sparsity problem brought by existing microorganism-drug association data set, greatly improves the prediction accuracy of microorganism-drug association effect.

This invention discloses a multi-scale, cross-domain interactive network for image tampering localization. Addressing the problems of existing methods that suppress semantic information to highlight forensic features, leading to the loss of key context and insufficient generalization ability, this invention constructs a multi-scale, cross-spatial, and noise-domain bidirectional interaction mechanism between semantic features and forensic features. Specifically, the network introduces a bidirectional cross-attention and adaptive gating fusion module, enabling high-level semantic information to guide the discovery of low-level tampering artifacts. Simultaneously, low-level forensic inconsistencies correct high-level semantic understanding, forming a closed-loop learning paradigm where semantics and forensic clues mutually reinforce each other. This method does not rely on specific target semantics and can effectively capture general patterns of image consistency violations, thus exhibiting excellent localization accuracy and robust generalization performance in both traditional editing and AI-generated tampering scenarios.