1124 results about "Differential privacy" patented technology

The invention provides a private weight self-adaptive heterogeneous data federated cooperative training method and system in the technical field of federated learning and privacy computing, and the method comprises the steps: S1, enabling each client to carry out the differential privacy operation on a local data set based on a private weight, and obtaining a desensitized data set, encoding the desensitized data set through a heterogeneous data encoding model; s2, performing semantic alignment on each coding vector through a contrast learning model to obtain an aligned vector set; s3, training a local model through the alignment vector set, generating a local gradient, extracting local model parameters, and uploading the privacy weight, the local gradient and local difference parameters to a server; and S4, the server trains the global model based on the local difference parameter and the global gradient, extracts the global model parameter and issues the global model parameter to each client for training. The method has the advantages that the compatibility, the flexibility and the efficiency of heterogeneous data federation cooperative training are greatly improved.

The invention discloses an accounting data checking method and system based on artificial intelligence, and the method comprises the steps: extracting multi-modal accounting data from a distributed tax data source through a federated learning framework, carrying out the anonymization aggregation of the data through a differential privacy technology, and generating a privacy-protected joint feature vector; inputting the joint feature vector into a causal reasoning model, identifying an abnormal fluctuation mode in the accounting data through anti-fact analysis, and outputting an abnormal index set with causal association; performing traceability reasoning on the abnormal index set by using a dynamic time sequence knowledge graph, generating a cross-cycle risk conduction path, and positioning a risk source entity; and generating an explainable inspection decision tree based on the risk source entity, dynamically adjusting an early warning threshold through adaptive threshold optimization, and outputting a graded early warning signal and a targeted inspection scheme. According to the embodiment of the invention, the accuracy, interpretability and risk traceability of distributed tax inspection can be improved.

The invention discloses a big data privacy protection modeling method and system based on federated learning and a block chain, and relates to the technical field of privacy protection and joint modeling. According to the method, homomorphic encryption, differential privacy, federated learning, secure multi-party computing and block chain technologies are fused, big data privacy protection and joint modeling are realized, encryption and dimensionality reduction are performed on original data through homomorphic encryption and differential privacy, an encrypted training sample of secure privacy is generated, a local model is trained on an encrypted data set through federated learning, and a big data privacy protection result is obtained. The method comprises the following steps: calculating aggregation parameters by using security multiple parties, constructing a verification network in combination with a block chain, ensuring credibility and integrity of model training, and finally, adding noise optimization performance for a global model by using differential privacy, testing generalization ability through cross validation, and determining a deployable privacy protection joint learning model, thereby breaking traditional data islands, promoting cross-mechanism data cooperation, and improving the privacy protection performance. Big data values are released, and data protection regulations and privacy requirements are met.

The invention discloses a medical data privacy protection method and device, and relates to the field of medical data security. The method comprises the following steps: acquiring medical data feature information of a plurality of medical institutions to form a medical data feature map; constructing a hierarchical federal learning framework based on the medical data characteristic spectrum; performing parameter aggregation optimization through a convolution weighting method of a recursive context guide network; a dynamic differential privacy budget allocation strategy is combined with a multi-first-choice Lambda weighted list DPO technology to optimize a noise injection process, and adaptive differential privacy protection is realized; and in combination with a homomorphic encryption technology, performing multi-party calculation through a security aggregation protocol to obtain a medical data privacy protection analysis system. According to the method, efficient cooperative analysis is carried out while the privacy of the medical data is protected, the problems that in the prior art, simple parameter exchange may cause model inversion attack, special optimization for medical scenes is lacked, and a supervision and auditing mechanism is imperfect are solved, and the utilization efficiency and safety of the medical data are improved.

The invention discloses an urban multi-mode digital twin data asset security management method, and relates to the technical field of urban digital twin and data security, and the method comprises the steps: firstly building a unified access gateway, converging multi-department and multi-type data, and generating sensitivity annotation and privacy level metadata; performing differential privacy or desensitization and dynamically calibrating parameters at the local part of the data provider according to the labels; bookkeeping access, authorization and training on a permission alliance chain, and performing fine-grained permission control; through on-chain authorization, each node only uploads disturbed and encrypted model update, and a global model is obtained after secure aggregation; the governance layer continuously monitors logs, budget and model indexes and triggers strategy closed-loop regulation and control; and multi-agent reinforcement learning of a privacy agent and a performance agent is further introduced, a unified multi-target optimizer is formed through distillation, privacy, federation and permission strategies are refreshed online, and collaborative analysis and compliance sharing operation without data out of a domain are realized.

The invention provides a multi-institution medical data collaborative modeling and intelligent generation method based on privacy protection, which relates to the technical field of privacy protection, and comprises the following steps: solving the problem of feature heterogeneity through a cross-institution feature semantic mapping graph, constructing a hybrid coding data structure to carry out hierarchical privacy protection, and carrying out multi-institution medical data collaborative modeling and intelligent generation. A homomorphic encryption mechanism is utilized to safely fuse multi-party local feature representations to generate a global representation, and differential privacy budget is allocated based on feature sensitivity. According to the method, multi-mechanism data security collaborative modeling is realized, and the model performance is improved while the data privacy security is guaranteed.

The invention belongs to the technical field of cross-border data security, and provides a cross-border personal data transmission security detection method and system based on differential privacy, and the method comprises the steps of multi-dimensional situation awareness, privacy budget allocation, interpretability and audit, and execution and response. By synchronously sensing technical environment risks, multi-national regulation constraints and multi-modal data features, comprehensive control of the cross-border transmission security situation is realized, and one-sidedness of single-dimensional sensing is avoided; privacy budget is dynamically allocated under multiple constraints based on a deep Q network algorithm, so that the protection intensity of high-sensitivity data is ensured, and excessive loss of effectiveness of low-sensitivity data is avoided; an interpretable report is generated in combination with causal analysis, full-process evidence storage is realized by relying on an alliance chain, the problems of decision black box and audit tracing are solved, cross-border supervision requirements are met, and the balance of cross-border transmission in the aspects of safety, compliance and availability is improved.

The invention discloses an energy storage battery health state evaluation method and system based on federated learning, and relates to the technical field of energy storage batteries. The method comprises the following steps: training a local health state evaluation model based on a federal loss function containing physical prior constraints at each client device, and enhancing sparse working condition data by adopting a local generation model to generate local model update; in the central server, performing value-guided heterogeneous aggregation processing, performing weighted aggregation on local model update submitted by the client, and generating a global health state assessment model; in the central server, executing digital twinborn consistency calibration, and performing post-aggregation fine tuning on the global model by using a reference signal generated by a cloud digital twinborn body; and communicating between the client equipment and the central server by adopting an event triggering and gradient sparse quantization compression mechanism, and deploying an adaptive differential privacy policy based on gradient inversion auditing. According to the method, the problems of high data privacy leakage risk, low model precision under heterogeneous data and high communication overhead in the prior art are solved.

The embodiment of the invention discloses a dynamic metadata sensing and self-adaptive mapping method and system, and the method comprises the following steps: deploying a lightweight probe to be agented to a plurality of heterogeneous source database systems, and building a trusted communication channel between the probe and a coordination center through equipment fingerprint generation and security authentication; a database log event is captured in real time, log analysis is carried out through an FPGA module, the change of a field structure or a semantic level is identified, and structured event data is generated; performing context embedding representation on the captured event based on a semantic understanding model, performing strategy reasoning in combination with a historical strategy and a knowledge graph, and generating a matched mapping strategy; loading and executing the mapping strategy in a trusted execution environment (TEE) to realize data conversion, cleaning and complementation, and performing desensitization processing on sensitive data through a differential privacy mechanism; and collecting performance indexes and abnormal logs of a strategy execution result, continuously training the model through a closed-loop optimization mechanism, and correcting an unreasonable strategy or updating a rule base.

The invention relates to the technical field of computers and information processing, in particular to a block chain-based cross-industry logistics data trusted sharing method and system, and the method comprises the steps: S1, carrying out the cross-industry semantic alignment; s2, performing field-level ciphertext access; s3, zero knowledge permission verification; s4, federal learning and on-chain coordination are carried out; s5, carrying out privacy budget and service linkage; s6, variable structure consensus; step S7, two-way auditing tracing is carried out; and S8, carrying out consistency challenge. Semantic alignment and chaining commitment are achieved through industry ontology, visibility is controlled through attribute-based field-level encryption and zero-knowledge verification, and commitment-driven median anchoring, symmetric truncation, quality * sample size weighted federal aggregation and on-chain differential privacy budget linkage are combined. And an end-to-end trusted sharing closed loop is constructed in cooperation with PBFT scalable consensus of VRF grouping, forward security signature and Merkle bidirectional auditing and challenge-response verification.

The invention discloses a virtual power plant capacity configuration and regulation operation optimization method, which comprises the following steps: constructing an aggregation model and physically consistent digital twinning, establishing a linearized power distribution network model containing voltage and power flow constraints, and depicting resource efficiency and time delay characteristics; forming a time-varying uncertainty set through quantile calibration and set drift constraint; establishing a capacity-operation joint double-layer optimization model, realizing capacity configuration under the constraint of the whole life cycle cost, and obtaining a rolling scheduling strategy through distributed robust optimization; before issuing, control shielding and formalized constraint are adopted to ensure the safety of the power grid; a multi-variety collaborative quotation is generated on the market side through opportunity constraint and risk measurement; hierarchical adaptive re-optimization is realized based on a trigger criterion, and a twin model is continuously calibrated by using a hardware-in-the-loop experiment; model updating is realized by adopting federated learning and differential privacy; executing degradation control under an abnormal condition, and performing smooth rollback after recovery; and finally, the capacity and operation parameters are evaluated and corrected through performance and service life linkage.

The invention discloses a dynamic privacy protection closed-loop control method and system for mobile edge crowd sensing. According to the method, information related to a communication channel and data content is obtained through risk perception, and a multi-dimensional comprehensive privacy risk score is calculated; then, dynamically determining a value of a differential privacy budget epsilon according to the risk score by utilizing a reinforcement learning algorithm in a strategy decision-making stage; then, performing noise addition disturbance on crowd sensing data according to the epsilon in a privacy protection execution stage so as to realize a preset privacy protection level; then, the privacy leakage risk and the data validity of the disturbed data are evaluated in the effect verification stage, and an evaluation result is obtained; and finally, privacy control strategy parameters are adjusted according to an evaluation result in a feedback optimization stage, and feedback is applied to a risk perception and decision process of a next cycle, so that closed-loop control is formed, and a privacy protection effect is continuously optimized. According to the method, the user data privacy security can be improved, and the data availability and the energy consumption efficiency are both considered.

The invention discloses a federal large model knowledge collaborative training method supporting multi-modal heterogeneous clients, which comprises the following steps: each client receives a model initialization parameter issued by a central server, and applies adaptive differential privacy noise to independently train a heterogeneous lightweight model based on local private data; updating the model to which the noise is applied and uploading a modal identifier of the model to a central server side; after model updating and modal identification of each client are received, based on a modal perception weighted consensus fusion mechanism, knowledge of each client is fused to update a global large model; and the central server side issues the updated presentation layer parameters of the global large model to the client side for initialization of the next round of local training. According to the method, on the premise that a public data set or specific task setting is not needed, comprehensive compatibility of data isomerism, client dynamic participation, model diversity and privacy protection requirements is achieved, and the adaptability, stability and knowledge utilization efficiency of large model federation training are remarkably improved.

The embodiment of the invention discloses a safety monitoring method and system for building construction, and the method comprises the steps: collecting the original video data of a construction site, carrying out the denoising, illumination correction and frame rate adjustment, and outputting a standardized video data stream; extracting attitude features of the constructors and representing the attitude features as a feature matrix to form an attitude feature matrix set containing spatio-temporal information; the sensitivity and correlation of the feature matrix are analyzed, calibration noise is added after dimension reduction, and feature data conforming to differential privacy are generated; and extracting data advanced representation by utilizing a pre-training model, and completing behavior classification, dangerous area judgment and safety violation detection. And the violation risk is evaluated in combination with the risk level of the construction area, graded early warning is generated, and meanwhile violation information is recorded to form a traceable management mechanism. According to the embodiment of the invention, accurate and efficient safety violation behavior detection is realized, and intelligent technical support is provided for safety management of building construction.

The invention discloses a dynamic portrait construction method and system fusing a large model and user behavior data, belongs to the technical field of artificial intelligence and big data analysis, and aims to solve the problems of insufficient real-time performance, difficulty in multi-source data integration and high privacy risk in the traditional technology. The method comprises the following steps: collecting basic attributes, behavior sequences and unstructured data by burying points, and processing through an Apache Flink session window and a dynamic watermark; multi-modal feature extraction (discrete feature embedding, bidirectional LSTM coding behavior sequence and BERT coding text) is carried out, and joint embedding is generated through cross-modal contrast learning; generating three types of labels, namely a static label (rule engine), a dynamic label (1.3 B parameter quantity Nano-vLLM) and a predictive label (XGBoost), and dynamically adjusting weights; and realizing global model updating through federated learning and differential privacy. The system comprises a data acquisition layer, a feature extraction layer and a label generation updating layer. The real-time performance and accuracy of the portrait are improved, the privacy of the user is protected, and the commercial value in e-commerce, finance and other scenes is remarkable.

The invention discloses an Internet of Things data storage method and device based on a block chain and a medium, and relates to the technical field of data storage, and the method comprises the steps: collecting an Internet of Things terminal data set, carrying out the noise disturbance and differential aggregation, and generating privacy protection data; performing hash coding on the privacy protection data, and outputting a tamper-proof to-be-stored data packet; performing security level encryption on the tamper-proof to-be-stored data packet by using a homomorphic encryption algorithm to form an anti-attack ciphertext vector; and carrying out data fragmentation and storage node mapping on the attack-resisting ciphertext vector to obtain a fragmentation storage index. According to the invention, through a dual privacy protection mechanism of homomorphic encryption and differential privacy, the privacy protection strength in the data storage process is improved. And meanwhile, safe and efficient multi-chain collaborative storage in a high-concurrency scene is realized through the constructed hierarchical cross-chain collaborative model.

The invention relates to the technical field of tax data security, and discloses a data sharing privacy protection platform responding to tax digital requirements. The platform comprises a tax data privacy perception layer, a sensitive information positioning mapping layer, a cross-source data security fusion layer, a compliance sharing decision-making layer and an execution state feedback layer. The tax data privacy perception layer obtains multi-source tax information based on a dynamic differential privacy mechanism, generates an initial privacy risk heat map and segments a sensitive area; the sensitive information positioning mapping layer generates a data internal privacy distribution stereogram through a security perception array and time sequence analysis; the cross-source data security fusion layer establishes a space-time mapping relation and generates a privacy disclosure risk score; the compliance sharing decision-making layer converts the score into an executable sharing instruction; and the execution state feedback layer monitors data change, generates an evaluation index and dynamically adjusts a privacy protection strategy. According to the platform, the privacy protection of the whole tax data sharing process is realized, and the accuracy and dynamism of risk identification and strategy adaptation are improved.

The invention relates to the technical field of industrial internet-of-things data security, in particular to an edge calculation differential privacy industrial internet-of-things data desensitization verification system and method.According to the system and the method, edge nodes are divided in a three-dimensional mode according to resource capacity, function positioning and privacy requirements, differential differential privacy parameters are formulated in combination with data attributes and leakage influences, and the safety of the industrial internet-of-things data is improved. And a dynamic adaptive grid is matched to realize hierarchical alignment, scene binding and elastic reconstruction, so that a cross-hierarchical privacy risk is avoided, and verification logic is simplified. Noise intensity is dynamically corrected by integrating multi-dimensional factors based on grids, lightweight, medium and deep hierarchical desensitization is designed for three types of nodes, and privacy protection and data availability are balanced. A privacy security and data availability two-dimension, terminal-gateway-area three-level verification chain is constructed, the desensitization effect is comprehensively evaluated, closed-loop iteration is achieved through hierarchical judgment and accurate adjustment, and industrial scene requirements are efficiently met.

The invention discloses an AI intelligent decision support method and system oriented to multiple scenes of a hospital and a doctor, and relates to the technical field of medical artificial intelligence, and the method comprises the steps: collecting multi-modal medical data, extracting a standardized feature vector, evaluating the feature sensitivity based on mutual information, and only injecting differential privacy noise into a high-sensitivity dimension to generate noise adding feature representation. And constructing a dual-path embedding model in a local training process, and respectively outputting a task prediction result and a federal feedback embedding abstract. Each participant uploads a local model parameter and an embedded abstract to the federated coordination end, and the coordination end generates a competition feedback signal, guides a next round of embedded optimization, calculates an aggregation weight based on a semantic offset degree, executes weighted aggregation to obtain a global fusion model, and completes reasoning output. On the premise of not leaking original data, cross-mechanism intelligent decision-making collaboration is realized, and the task accuracy and the feedback alignment capability are improved.

The invention discloses an intelligent traditional Chinese medicine decoction control method based on the Internet of Things and a storage medium, and relates to the technical field of traditional Chinese medicine preparation equipment and process control. A target state vector is generated through a prediction model deployed on a terminal side, and a benchmark reference is provided for off-line control; the fine adjustment of decoction parameters is realized by controlling an increment mapping function, and the stability of the control process is ensured by a rate limitation and amplitude saturation mechanism; due to introduction of the safety boundary condition, abnormal working conditions such as pressure overrun and too low liquid level are effectively prevented; after the network is recovered, an incremental training mechanism of differential privacy protection is adopted, so that off-line data is fully utilized, and the risk of privacy disclosure is avoided; a dynamic weight distribution strategy in cloud federation aggregation ensures the leading effect of high-quality data on model optimization, and a multi-stage verification mechanism ensures the reliability of model updating.

The invention discloses an intelligent monitoring method for a hydrogeological environment, and the method comprises the steps: 1, constructing a space-air-ground integrated sensing network, and dynamically adjusting the monitoring frequency through a self-adaptive sampling strategy; 2, developing cross-platform data middleware, realizing conversion from multi-protocol data to a WaterML 2.0 standard, constructing an anomaly detection model in combination with a deep residual network, and correcting regional parameter deviation; 3, designing a hybrid fusion architecture, processing time sequence data by using an improved Transform at a bottom layer to capture long-term dependence, constructing a graph neural network at a top layer to fuse spatial connectivity, and outputting a prediction result with a confidence interval through an attention mechanism and a Bayesian network; step 4, establishing an edge-cloud hierarchical computing system, deploying lightweight model real-time early warning at an edge end, integrating complex models such as 3D geological modeling at the cloud end to generate a regional report, and adopting federal learning cooperative training and embedding differential privacy protection; and 5, developing a three-dimensional geological visualization platform and an intelligent decision-making system.

The invention discloses a differential privacy data set distillation method and system based on image generation data, and belongs to the technical field of data privacy protection and machine learning. The method comprises the following steps: firstly, synthesizing a synthetic data set meeting Gaussian differential privacy, training a feature extractor on the synthetic data set, and finely adjusting an expert model by using original data differential privacy; multiple rounds of iterative optimization are carried out on the distillation data set initialized according to the classes, wherein a feature extractor is randomly selected according to the classes in each round to align the features of the original data added with the noise and the distillation data, and an expert model is utilized to align the semantics of the distillation data hard labels and the semantics of the synthetic data soft labels; and finally outputting a distillation data set meeting the total privacy budget. According to the method, effective priori is provided by utilizing generated data, extra noise injection in the alignment process is reduced, convergence is accelerated, and better privacy protection and data availability balance compared with a previous method is realized under the same differential privacy budget.

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.

The present application relates to the field of federated learning and privacy protection, and particularly relates to a direction-aware local differential privacy federated learning method. The present application effectively solves the problem of serious decline in model performance caused by the destruction of gradient direction information by traditional differential privacy methods by adopting a direction-aware noise selection strategy based on an exponential mechanism locally on the client side. Specifically, the client first performs L2 norm clipping on the gradient to control the sensitivity, then generates K candidate noise vectors from a Gaussian distribution, evaluates the direction consistency of each candidate noise and the true gradient through cosine similarity, and selects the noise closest to the gradient direction in a probabilistic manner using the exponential mechanism, and finally adds the selected noise to the gradient to complete the perturbation. The method provides a strict epsilon-local differential privacy mathematical proof for the entire noise selection process, reduces the calculation time by about 80% compared to existing direction-aware methods, and achieves a triple balance of privacy protection, model utility and computational efficiency.

The invention discloses a lithium battery SOH and RUL prediction system and method based on multi-scale feature federation, and relates to the technical field of battery detection, the system comprises a data acquisition and feature extraction module, an encryption and uploading module, a model construction module, a recursion updating module and a work order generation module; through the multi-scale feature fusion formula, the time sequence features and the statistical features of the lithium battery are organically combined, the fusion feature vector is generated, degradation information of the lithium battery under different time scales and operation conditions is comprehensively captured, the health state of the lithium battery and the accuracy of prediction of the remaining service life of the lithium battery are improved, and the prediction accuracy of the remaining service life of the lithium battery is improved. And meanwhile, through dynamic Top-K sparse compression, 8-bit linear quantization and AES-256-GCM encryption technologies and in combination with a differential privacy protection mechanism, the data transmission quantity is reduced, the data security is enhanced, sensitive information leakage is prevented, and a reliable guarantee is provided for remote monitoring and predictive maintenance of the lithium battery.

The present invention relates to the field of privacy computation. Disclosed are a secure multi-party computation method and system based on a decision tree and privacy protection. In a data processing stage, the method uses a gradient-based one-sided sampling algorithm, so as to exclude most small-gradient samples, thus ensuring balance in accuracy while reducing the data volume; then, the method uses a histogram algorithm to construct a decision tree, thus reducing the memory consumption and increasing the computational speed; and in addition, the method also uses a differential privacy technology to perform encryption and perturbation processing on parameters of local models, so as to ensure that individual privacy information is not leaked. The secure multi-party computation method based on a decision tree and privacy protection of the prevent invention can implement prediction tasks such as classification efficiently and reliably while protecting privacy.

The invention discloses a data security sharing method and system, and relates to the technical field of data sharing. The method comprises the steps that a data provider uploads original data, a privacy budget value is calculated through a differential privacy algorithm, noise is added to obtain noisy data, and the noisy data is processed through an anonymous privacy protection algorithm based on maximum dissimilarity degree clustering to obtain desensitized data; encrypting the desensitized data by adopting a block chain decentralization-based ciphertext policy attribute-based encryption method, obtaining an encrypted data ciphertext, uploading the encrypted data ciphertext to a cloud server, and obtaining a content addressing hash value; encrypting the hash value through an elliptic curve encryption algorithm, and storing the encrypted hash value to a block chain account book; and when an access request of a data requester is received, the cloud server verifies the authority by using a non-interactive zero-knowledge proof protocol and returns an encrypted hash value after passing the verification, and the requester decrypts to obtain the hash value and decrypts the encrypted ciphertext to obtain desensitized data, thereby realizing data sharing.

The invention discloses a federal learning-based cross-device brain grain model training system. The system comprises a terminal device layer, an edge coordination layer and a central server layer, the terminal equipment layer is responsible for collecting original electroencephalogram data and executing local lightweight training; the edge coordination layer is used for aggregating model updates of a plurality of terminal devices in the jurisdiction to generate a regionalized sub-model; according to the central server layer, a central server performs secondary optimization on each regional model to generate a global unified model; the encrypted global model and strategy configuration are pushed to all edge coordination layers; and the edge coordination layer pushes the optimized model and strategy to terminal equipment. Through a three-layer federated architecture and multi-dimensional optimization, the system efficiency is remarkably improved: the terminal equipment only uploads an encrypted model increment, and source data protection is realized in combination with differential privacy and homomorphic encryption; federal transfer learning shortens the cold start time of new equipment.

The invention discloses a federal learning security training method and system based on differential privacy, belongs to the technical field of artificial intelligence, and aims to solve the technical problem of how to realize data privacy protection and model precision balance in cross-mechanism model training. Comprising the following steps: constructing a distributed structure comprising a plurality of clients and a server; each client constructs a noise variance calculation model by taking the gradient feature, the privacy budget coefficient, the data sensitivity level coefficient and the training stage coefficient as calculation parameters after each round of local training is finished, and Gaussian noise is calculated according to the noise variance; verifying the noise variance of each round, and adjusting the calculation parameters of the noise variance based on the comparison result of the accumulated budget consumption and the global privacy budget; dynamically allocating the data sensitivity level of the local training data and the budget allocation proportion of the client; and the server performs secure aggregation through a secret sharing algorithm, and distributes the updated global model parameters to each client.

The invention relates to the technical field of intelligent vocational ability analysis, and discloses a dynamic planning method for a vocational ability genetic map based on multi-modal data and federal learning. The method comprises the steps that multi-source heterogeneous data in occupational activities is collected, and a multi-modal data set is formed through data cleaning and standardization processing; a graph neural network technology is utilized to construct a correlation topology between capability elements, spatio-temporal context information is fused, a dynamically evolved vocational capability gene map is established, and the accuracy and interpretability of vocational capability analysis are improved. Multi-node cooperative training is realized by adopting a federated learning framework, data privacy security is ensured in combination with a differential privacy technology, and the problem of data islands is solved. A capability evolution path is simulated based on a three-dimensional visualization and digital twinning technology, an occupational development scheme is generated through an intelligent recommendation algorithm, and a real-time feedback optimization mechanism is established. According to the method, dynamic, visual and personalized vocational ability analysis is realized, and accurate decision support is provided for vocational planning.