324 results about "Learning architecture" patented technology

The invention discloses a marine environment real-time monitoring and early warning system based on machine learning, and relates to the technical field of machine learning. Comprising the steps that an ocean multi-source sensing module collects ocean environment data in real time through a sensor and a combined collection scheme; the multi-source feature extraction module performs time domain, change rate and frequency domain feature analysis on the data to construct a unified multi-dimensional feature vector; the multi-model fusion prediction module outputs a marine environment state vector through dynamic weighting and deviation correction based on a parallel learning architecture of a deep neural network, a long-short-term memory network and a one-dimensional convolutional neural network; and the ocean risk identification and early warning module generates graded and classified early warning information through double study and judgment of a sea condition classifier and an abnormal event detector. According to the method, comprehensive acquisition, deep feature mining, high-precision prediction and accurate early warning of marine environment data are realized, the problems of low prediction precision, risk identification lag and the like in the prior art are effectively solved, and reliable guarantee is provided for marine operation safety.

Artificially intelligent systems and methods for financial coaching provide personalized, fiduciary-compliant financial guidance through advanced machine learning architectures with measurable performance criteria. The systems implement privacy-preserving processing pipelines that detect personally identifiable information using multi-layered pattern recognition including regular expressions for formatted data sequences, named entity recognition with confidence thresholds above 0.85, and contextual analysis algorithms. A multi-step artificial intelligence processing workflow includes automated language detection, emotional tone classification with confidence scoring, financial profile transformation using predefined templates, context-aware question rephrasing, and semantic similarity matching employing vector embeddings with financial domain vocabulary weighting applying multiplier values between 1.3-2.0. Specialized training methodologies expand datasets through mathematical transformation functions utilizing statistical standard deviations with incremental variations between 0.5-2.0. Mood-based escalation logic automatically transfers users to human advisors when emotional indicators exceed confidence thresholds above 0.8. The systems maintain response times below 5 seconds while providing regulatory compliance through curated content sources and predefined fiduciary instruction parameters.

The invention relates to an unmanned aerial vehicle group-oriented sensing communication integrated network multi-resource joint scheduling method, and belongs to the technical field of wireless communication, and the method comprises the steps: building a system model of multiple UAV-ISAC tasks, and defining a joint optimization problem; extracting spatio-temporal features from the dynamic heterogeneous graph in which the unmanned aerial vehicle, the user and the sensing target are abstracted as nodes and the relationship is abstracted as edges; taking the features as input, and adopting a layered multi-agent reinforcement learning architecture to solve the joint optimization problem on line; in the architecture, resource allocation and trajectory planning actions are generated through cooperation of a central Actor and all unmanned aerial vehicle Actors, and system performance is evaluated by a central Critic; constructing a multi-target weighted reward function, stabilizing a training process by combining experience playback and a Mini-batch sampling mechanism, and updating network parameters in parallel; and obtaining an optimal resource allocation and unmanned aerial vehicle trajectory strategy through training. The sensing performance is improved, the communication quality is guaranteed, and the defects in the aspect of dynamic resource scheduling in the prior art are overcome.

The invention discloses a protocol conversion and protocol self-learning method and system for transformer area optical storage and charging cooperation, and the method comprises the steps: constructing a digital twinborn body of a transformer area optical storage and charging system, carrying out the parallel operation of a protocol agent and a protocol agent in a virtual environment, and achieving the cooperative training of the protocol agent and the protocol agent through a hierarchical reinforcement learning architecture, the protocol agent is responsible for learning a dynamic priority scheduling and compression strategy for heterogeneous protocol messages such as Modbus, CAN and IEC 104, the protocol agent is responsible for learning a power balance and voltage stability control strategy based on photovoltaic output, energy storage SOC and charging load, and the two agents realize cross-domain collaborative optimization through a reward function mutual coupling mechanism. And finally, safely deploying the collaborative strategy obtained by training to the edge control equipment of the physical transformer area. According to the method, the problems of disjunction of protocol conversion and cooperative control, protocol strategy solidification, insufficient cross-domain cooperation and the like in the prior art are solved, and the operation efficiency and the self-adaptive capability of the transformer area optical storage and charging system are improved.

The invention discloses a flow field video generation method based on a policy value architecture and online physical exploration, and belongs to the technical field of crossing of artificial intelligence and computational fluid dynamics (CFD), and the method comprises the following steps: step 1, constructing an unsteady flow field multi-modal training data set, step 2, constructing a generative network system based on an Actor-Critic architecture, step 3, constructing an unsteady flow field multi-modal training data set, and step 4, constructing an unsteady flow field multi-modal training data set. Step 4, supervised fine tuning training is carried out in the first stage; step 5, online physical exploration of a generator is carried out in the second stage; step 6, feedback co-evolution of a physical encoder is carried out in the third stage; and step 7, reasoning generation of an unsteady flow field video is carried out. According to the method, a reinforcement learning architecture containing an Actor and a Critic is constructed, a physical equation is packaged into a digital environment, and a training strategy of basic supervision fine tuning, online physical exploration of a generator and coevolution feedback of an encoder is adopted.

A marine zooplankter identification method based on in-situ images and deep learning is used for processing a multi-stage series deep learning architecture constructed for all in-situ images, and comprises the following steps: quickly positioning and framing zooplankter individuals under a complex background based on a YOLOv8 skeleton construction model; performing pixel-level fine segmentation on individuals in the in-situ image by using a U-Net model, and extracting an accurate contour to obtain a high-quality individual image; and finally, inputting the image into a PlanktonNet network innovatively designed based on a ViT architecture, and through adaptive small-size slice embedding and introduction of a convolution word embedding layer, improving small-scale target feature extraction capability, and realizing high-precision and fine-grained classification of genera and species. The method has the advantages that target detection, semantic segmentation and recognition tasks are organically fused, the problems of low recognition efficiency and low automation degree in the prior art are effectively solved, and marine zooplankton can be quickly and accurately recognized from in-situ images on a large scale.

The invention discloses a data deep learning and intelligent analysis method based on an AI artificial intelligence technology, and relates to the technical field of basic AI models, and the method comprises the steps: employing a multi-modal data preprocessing module to carry out the expansion of small sample data through a generative model, and combining with meta-learning to extract prototype features, meanwhile, an epsilon-differential privacy budget is dynamically allocated based on the data sensitivity level so as to inject dynamic noise; establishing a layered federated learning architecture, training a model by local training nodes through a loss function containing a self-adaptive regularization item, and performing sparse processing and gradient disturbance before uploading parameters; the global aggregation node adopts a weighted federated average algorithm to aggregate parameters, and dynamically adjusts the communication frequency according to the loss convergence speed; and a target model is obtained through iterative training, and a decision interpretation report containing the attention thermodynamic diagram and the desensitization identifier is generated when a result is output. According to the method, the problems of small sample overfitting, data islands and privacy disclosure are effectively solved, and the accuracy and practicability of the model are improved.

A method includes receiving a video comprising one or more frames; executing a first machine learning model using the one or more frames of the video to generate a dynamic mask configured to track a predicted magnitude of attention that individuals will give to different portions of each of the one or more frames of the video during playback of the video, the dynamic mask comprising attention scores for individual portions of each of the one or more frames of the video; generating one or more attention metrics for the video based on an aggregation of attention scores for the individual portions of each of the one or more frames of the video; and generating a record identifying the one or more attention metrics for the video.

The invention relates to the technical field of computers, and discloses a fragmentation block chain federal learning method based on a large language model multi-agent, and the method comprises the steps: S1, initializing a client; s2, dynamic fragmentation scheduling and distribution; s3, generating and uploading local knowledge; s4, intelligent agent collaborative routing and knowledge acquisition; s5, knowledge fusion and model updating; and S6, repeatedly executing the steps S3 to S5 until the model converges or reaches a preset number of iterations. According to the invention, under a decentralized and fragmented federated learning architecture, the complex reasoning ability of a large language model and the autonomous cooperation mechanism of three multi-agent systems, namely a fragmented scheduling agent, a fragmented knowledge state agent and a global knowledge routing agent, are deeply fused; according to the mechanism, dynamic optimization of a bottom layer fragment structure and intelligent routing of high-value knowledge are achieved in an intelligent mode, and therefore the overall efficiency and model performance of a system under the condition of heterogeneous data and heterogeneous equipment are remarkably improved.

Disclosed are systems and methods including software processes for developing test cases for testing robustness of AI-based image-recognition models-under-test (MUTs) with respect to types of image variation transformations. The system may generate various types of robustness metrics for the MUT and output user-readable reports about the MUT's performance. The system trains machine-learning architectures to generate test cases including augmented images according to the types of image transformations, applies the IR MUTs, and then evaluates the image feature vector embeddings and predicted classification produced by the IR MUTs to determine the accuracy of the MUT with respect to each type of transformation.

The invention discloses a medical record large model privacy enhancement fine tuning method based on a federated learning framework. The method comprises the following steps: S01, constructing a three-level medical exclusive federated learning framework of a client, an edge node and a federated server; s02, the client performs privacy enhancement preprocessing on the medical record, generates a medical exclusive feature vector and adds differential privacy noise; s03, a three-level privacy enhancement mechanism of a data layer, a transmission layer and a training layer is designed, dynamic differential privacy noise injection is adopted in the data layer, homomorphic encryption and secure aggregation are adopted in the transmission layer, and gradient mask and federated distillation technologies are adopted in the training layer; s04, adopting a federal fine tuning strategy adaptive to medical characteristics; and S05, constructing privacy security verification, clinical compliance auditing and a model iterative optimization closed loop. Therefore, the privacy protection intensity is obviously improved, the medical data adaptability is optimized, the heterogeneous deployment flexibility is enhanced, the privacy and performance are dynamically balanced, the clinical compliance is guaranteed, and the cooperation efficiency is improved.

The embodiment of the invention provides a multi-level federal multi-mode large model privacy protection enhancement method and electronic equipment, and belongs to the technical field of intelligent traffic systems. According to the method, a'end-edge-cloud 'three-level federated learning architecture is constructed, firstly, local differential privacy disturbance is applied to local multi-mode traffic data at a traffic terminal node, and an end-side local model is trained; then aggregating a plurality of end side models on an edge server, and generating a personalized small model reflecting regional characteristics; then cooperatively training a plurality of personalized small models in a cloud center server to generate a global multi-modal large model; finally, knowledge of the global large model is fed back to a lower-level model through knowledge migration, and a continuously evolved iterative closed loop is formed. According to the method, privacy protection is carried out at a data source, so that original sensitive data is ensured not to go out of the local, the problems of data islands and privacy leakage in traffic large model collaborative training are effectively solved, and efficient and credible collaborative modeling is realized on the premise of ensuring data sovereignty.

The invention relates to the technical field of new energy intelligent vehicle energy management, and discloses an intelligent vehicle power supply energy management method based on deep learning, comprising the following steps: step A, collecting vehicle dynamic operation parameters, power battery electrochemical state parameters and external driving environment parameters, and constructing a multi-source heterogeneous data stream; and step B, inputting the multi-source heterogeneous data stream into a hierarchical deep learning architecture, wherein the hierarchical deep learning architecture comprises a convolution-long and short-term memory hybrid network for spatial-temporal feature extraction and an attention mechanism multi-task network for jointly estimating the state of charge and the state of health of the battery. Virtual rehearsal and safety threshold verification are carried out on a power distribution strategy through a digital twinborn model, potential risks are identified before an instruction is issued, increment retraining is triggered, a closed-loop feedback mechanism is formed, the defects that a traditional deep learning scheme is insufficient in safety and rigid in a vehicle-mounted scene are overcome, and the safety of the vehicle-mounted scene is improved. The system robustness is obviously improved; and the service life of the battery is prolonged.

The invention discloses a large model-oriented split privacy protection training method, which comprises the following steps that: a pre-trained large model is divided into a client side model bottom layer and a top layer and a cloud server side model middle layer along a network depth direction, a client only processes local private data, and a terminal cloud performs cooperative training through middle activation. The method is oriented to privacy protection of a large model, a privacy strategy adaptive to characteristics of the large model, application of a cross-model architecture, realization of accurate privacy-utility balance, noise injection of token importance perception, lightweight noise post-calibration, provision of privacy guarantee of theoretical constraints, a token-level differential privacy mechanism and parameterized adjustable privacy constraints. According to the method, deployment feasibility and system efficiency are optimized, extra calculation and storage overhead is low, an efficient parameter fine tuning technology is compatible, the communication traffic and calculation requirements are greatly reduced while the privacy protection effect is guaranteed, and efficient and practical privacy fine tuning of a large-scale model under a split learning architecture is supported.

The invention relates to the technical field of global navigation satellite system reflection, in particular to a GNSS-R sea surface wind speed inversion method, which comprises the following steps of: firstly, comprehensively considering multiple constraint conditions such as inversion error distribution characteristics, interval quantity, sample distribution uniformity and interval minimum sample requirements by utilizing the global optimal search characteristic of an SA algorithm; a multi-objective optimization function is constructed to preliminarily divide a wind speed interval, and then a GD algorithm is adopted to accurately adjust the boundary of the preliminarily divided interval. And constructing a special XGBoost prediction model for each wind speed interval, and finally, fusing wind speed prediction results of a plurality of sub-interval models through a stacked integrated learning architecture to obtain a wind speed inversion value. According to the method provided by the invention, the problem that the inversion error of a traditional single model in different wind speed ranges is extremely unstable can be effectively relieved, the inversion precision of the whole wind speed range is improved, and meanwhile, the problem that a real-time observation sample cannot correspond to a sub-model under an interval model in practical application is solved; and the application conversion of the GNSS-R sea surface wind speed inversion technology is promoted.

The invention provides a low-altitude visual analysis and illegal behavior recognition method for a port, and the method comprises the steps: collecting and carrying out the time-space alignment of a video stream through a multi-view high-definition camera, and extracting a standardized behavior event sequence through a deep neural network and a tracking algorithm; the events are sequenced according to time, behavior frequency, transition probability and co-occurrence modes are counted in a sliding time window, a dynamic feature vector is generated to drive an incremental graph neural network, and continuous embedding updating, confidence calculation and anomaly detection are performed on a behavior semantic graph; if structural abnormity is detected, a new behavior mode is identified through clustering analysis and causal reasoning assistance, and map self-adaptive adjustment is achieved after expert feedback verification; and finally, a composite semantic map is constructed by fusing conventional and abnormal behavior reasoning results through a double-channel learning architecture, port violation behaviors are accurately identified and pre-warned, and the scene understanding ability and the safety monitoring intelligent level are improved.

Systems and methods are disclosed for automated generation, validation, and deployment of infrastructure-as-code (IaC) using a multi-agentic artificial intelligence / machine learning (AI / ML) architecture. An extraction agent parses multimodal artifacts (e.g., diagrams and / or configuration data) to derive or generate infrastructure set-up parameters. A coding agent generates IaC units based on the extracted parameters. A validation agent determines the syntactic and semantic compliance of the generated IaC, classifying the IaC as executable, non-executable, or identifying corrective actions. A deployment agent transmits executable IaC to target computing environments and manages automated provisioning of the IaC in the target environments. In cases of validation failure, error indicators are provided to a user (e.g., via an agentic chat bot) for clarification or correction, enabling iterative refinement.

The invention discloses a road investigation design method and system based on a mobile network, and relates to the technical field of road investigation, and the method comprises the steps: dividing geological anomaly probability distribution data into dynamic grid units, constructing a reinforcement learning state space, and generating an optimal sampling path instruction in combination with a dual-network deep Q learning architecture; analyzing the optimal sampling path instruction into executable parameters, executing the executable parameters, collecting multi-source spatio-temporal data, establishing data association through a spatio-temporal hash algorithm, and generating a reconnaissance data set; and on the basis of the survey data set, through parameterized spline curve modeling, generating a candidate road design scheme, and in combination with a non-dominated sorting genetic algorithm, performing optimization to generate a three-dimensional road design scheme. According to the method, the exploration path is optimized by using the dynamic grid coding and the dual-network architecture, the high-risk area coverage and the moving efficiency are balanced, meanwhile, the method adapts to real-time geological changes in combination with priority experience playback, and the exploration reliability and the resource utilization rate are improved.

Mechanisms are provided for optimizing a deep learning (DL) computer model for homomorphic encryption (HE) workload processing. The mechanisms receive an original DL computer model architecture that is to be optimized for HE workload processing, and modifying the original DL computer model architecture by replacing a self-attention layer of the original DL computer model with an HE friendly self-attention layer that comprises a Power SoftMax function that does not have exponent terms, to thereby generate a modified DL computer model architecture. The mechanisms execute a machine learning training of the modified DL computer model architecture, approximate one or more elements of the Power SoftMax function with polynomials to generate a trained HE optimized DL computer model, and output the trained HE optimized DL computer model for execution on HE workloads.

The invention discloses a phosphorus chemical industry production equipment on-line state monitoring and intelligent diagnosis system based on multi-parameter fusion, and relates to the technical field of chemical industry equipment intelligent monitoring and diagnosis. Functional modules such as data input and preprocessing, equipment wear evaluation, real-time data analysis, fault prediction, personnel risk analysis and comprehensive diagnosis are deeply integrated, and the system accurately adapts to non-stationary characteristics of the phosphorus chemical process by dynamically adjusting a time window and a self-attention weight. The comprehensive diagnosis module innovatively constructs a causal diagram based on self-attention features, generates a structured diagnosis report by using a Transform decoder, and optimizes the diagnosis probability in combination with Bayesian reasoning, so that the accuracy, comprehensiveness and intelligence level of monitoring diagnosis are improved, accurate early warning can be realized, deep insight and optimization operation and maintenance decisions can be provided, and the system has a wide application prospect. And the production risk and economic loss are effectively reduced.

The invention discloses a screen candid shooting prevention monitoring method and system based on multi-mode model fusion, and relates to screen privacy protection.The method comprises the steps that firstly, multi-source data from a video mode, a wireless positioning mode, an environment light intensity mode, an audio mode and a peripheral behavior mode are acquired through a data acquisition module; and the data enhancement module is used for enhancing the source data. Then, feature extraction is carried out on the enhanced multi-modal data, and corresponding multi-modal feature representation is obtained; and inhibiting inter-modal redundant information through an adaptive feature enhancement module, highlighting significant features related to candid photographing behaviors, and generating optimized fusion features. And on the basis of the optimized fusion features, a multi-task learning architecture module is adopted to simultaneously predict a device category, a user behavior category and whether a secret photographing behavior exists or not. According to the method, multi-modal data can be fully fused, the accuracy and robustness of screen candid shooting prevention detection are improved, and the method is suitable for real-time security monitoring of multiple scenes and multiple types of terminals.

The invention discloses a tunnel construction robot path planning method and system based on deep reinforcement learning, and relates to the technical field of tunnel construction automation and robot navigation, and the method comprises the steps: pre-defining a state space and an action space of robot path planning in a tunnel construction environment, and building a dynamic reward function; constructing a deep reinforcement learning architecture based on a lightweight action network and a comment network, and training an action-comment network based on state-action pair data in an experience pool by taking the sum of an accumulated reward and an entropy regularization item of a maximization strategy as a target; and on the basis of the current state information of the tunnel construction robot, the optimal moving action of the next step is obtained by utilizing the trained action network, and path planning of the robot is realized. According to the method, the deep reinforcement learning technology is applied to the tunnel construction scene, and the autonomous decision-making ability and the path planning precision of the robot in the dynamic construction environment are optimized and improved in combination with real-time data acquisition and processing of the complex tunnel environment.

Various embodiments of the present disclosure provide machine learning architectures and data processing techniques for improving computer-based text comprehension. The techniques include generating, using a trained classifier model, target classification probabilities for labelled text-based objects from a testing portion of a labelled training dataset and identifying predictive text-based objects from the labelled text-based objects based on the target classification probabilities. The techniques include applying a staged prompting mechanism with a generative extraction model to identify a target set of explanatory text segments from the predictive text-based objects that may be clustered into semantic segment clusters. The techniques include generating explanatory summary segments respectively corresponding to the semantic segment clusters and generating a target classification signature based on a plurality of terms from the one or more explanatory summary segments.

The invention is suitable for the technical field of multi-modal data fusion, and provides a multi-modal data fusion deep learning geological map intelligent prediction method, which comprises the following steps: acquiring multi-modal geological data, constructing a heterogeneous multi-modal feature extraction network, and respectively extracting high-dimensional feature representation of each modal data by adopting a special encoder; designing a multi-modal dynamic fusion module based on a multi-head attention mechanism; establishing a deep learning architecture of a geological knowledge guiding mechanism, and embedding geological rules and priori knowledge into a model training process through a constraint loss function; implementing a multi-task collaborative learning framework, and synchronously completing lithology classification, structure identification and mineralization potential prediction tasks; outputting reliability evaluation of a prediction result by adopting an uncertainty quantification system of a Bayesian deep learning framework; deploying an intelligent active learning system; according to the multi-modal dynamic fusion method based on the multi-head attention mechanism, the problem of depth feature interaction and complementation among heterogeneous geological data is effectively solved.

A machine-learned architecture may use sensor data, data derived from the sensor data, and / or to determine a grid indicating objects that are occluded to one or more sensors of a vehicle and velocities associated therewith. This grid may be used to generate a candidate object detection associated with an occluded object and / or an object track associated with the occluded object. The candidate object detection and / or object track may be used to determine a predicted trajectory of the occluded object and, in cases where the occluded object is affirmed by a perception component of the vehicle to be a true positive, may be used to initialize a track associated with the newly disoccluded object.

The invention discloses a main and distribution network collaborative analysis method, system and device and a storage medium, and belongs to the technical field of power system analysis and calculation. The method comprises the following steps: abstracting a main power distribution network integrated model into a heterogeneous graph and generating a fusion node feature matrix; a time-space diagram neural network model is constructed and trained, the model ensures that a prediction result conforms to a power system physical law by introducing physical information neural network constraints, and multi-task parallel prediction is completed by adopting a diagram self-attention pooling and multi-task learning architecture; performing cross-domain interaction verification on a collaborative analysis result output by the model, wherein the cross-domain interaction verification comprises boundary consistency verification and energy / power balance verification; and dynamically adjusting calculation granularity or model parameters according to a verification result, and accelerating a calculation process based on a linear self-attention mechanism. According to the method, the problems of model splitting and inconsistent calculation results caused by independent analysis of the traditional main and distribution networks are solved, and global consistent, physically credible, efficient and collaborative power grid analysis is realized.

The invention provides a community electric vehicle charging load prediction method and system based on machine learning. The method comprises the steps of collecting historical four-dimensional data related to electric vehicle charging in a community; performing fusion processing on the data to obtain a spatial-temporal characteristic matrix; constructing a load prediction model corresponding to each charging pile in the community, and training the load prediction model of each charging pile by using the spatial-temporal characteristic matrix; each charging pile encrypts the parameters of the trained load prediction model, uploads the encrypted parameters to a central server, and aggregates the model parameters under an improved federated learning architecture to obtain a global load prediction model; and collecting real-time four-dimensional data related to charging of the electric vehicles in the community, performing prediction based on the global load prediction model, obtaining a predicted charging load in a specified time period in the future, and adjusting the charging electricity price according to the predicted charging load. The method can improve the generalization capability of the prediction model, and improves the prediction precision of the community electric vehicle charging load.

The invention relates to a radar resource intelligent scheduling method and device based on multi-agent reinforcement learning. The method comprises the steps of obtaining a target priority evaluation matrix based on a multi-dimensional comprehensive state vector containing a self state, a target state, a time sequence evolution characteristic and agent collaborative information, a training strategy and a value network by adopting a reinforcement learning architecture of centralized training and distributed execution, and obtaining a target priority evaluation result through a greedy allocation and dynamic replacement strategy. And converting into a beam and residence time allocation scheme meeting hard constraints. And fusing observation data through a state estimation algorithm to obtain a tracking result and a performance index, performing feedback iteration optimization on the network in combination with a multi-target weighted reward function, and finally outputting a real-time scheduling result by utilizing the trained network. By adopting the method, the distribution can be dynamically adjusted under the scene that the target density fluctuates and the resources are temporarily limited, the tracking continuity is ensured, and the risk of missing tracking is greatly reduced.

The embodiment of the invention relates to the technical field of visual feedback, and provides a self-adaptive visual feedback system and method for ensuring material forming quality, the system comprises a front-end framework and a rear-end framework which work cooperatively, the front-end framework comprises a sensing module, a feature extraction module, an error compensation module and an execution control module, the back-end architecture comprises a decision module, a twin simulation module and an evolution module; the system further comprises a quality analysis module. According to the embodiment of the invention, environmental interference is overcome through active vision and event camera collaborative perception, long-term precision is maintained through online calibration compensation, quality quantitative evaluation is realized by combining three-dimensional shape real-time reconstruction and a deep learning method, a digital twin technology is introduced to carry out strategy security verification, and continuous evolution of the system is realized by means of a federated learning architecture. The method is not only suitable for the welding process, but also can effectively meet the high-quality control requirements of various material forming processes such as spraying, cladding and surfacing.

The invention provides a federated learning and privacy calculation-based cross-border earthquake risk assessment method, which solves the fundamental contradiction between data sovereignty and privacy security in cross-border data sharing by constructing a federated learning architecture fusing hierarchical desensitization and hybrid encryption. Through the multi-source data cooperative training and CNN-LSTM mixed disaster chain model, the problem that the assessment precision is insufficient due to the fact that a single country model cannot describe a complex cross-border disaster chain is solved; finally, a cross-border disaster scene is constructed, and the defect that a static evaluation tool is disjointed with an actual combat is overcome through a multi-agent dynamic policy simulation system; according to the method, the security collaboration under the condition that the data is not out of the scene is realized, the multi-country collaboration participation rate is improved by 60%, the risk assessment accuracy rate is improved to 92% or above, the policy assessment period is shortened from several months to several hours, a quantifiable scientific basis is provided for cross-border emergency decision making, and the joint prevention and joint control capability of regional earthquake disasters is effectively improved.