385 results about "Data privacy protection" patented technology

The invention provides a data privacy protection method for a data governance system, and belongs to the technical field of data governance, and the method comprises the steps: carrying out the cross verification of multi-source feature data and the unique identification information of an object collected on site, generating an original data set, carrying out the sensitive information recognition and grading, and generating the preprocessing data with a sensitive grade label; core identification information in the preprocessed data is disassembled to generate standardized desensitized data conforming to privacy protection, a bidirectional encryption mapping relation is established between object unique identification information collected on site and the standardized desensitized data, an encryption index is formed, the encryption index and preset multi-dimensional compliance data are fused, and a data fusion result is obtained; generating standardized fusion data; and based on the access token, generating a differential authorization data set divided according to permission granularity, performing privacy disclosure risk assessment, generating a risk level, performing privacy processing on the risk level, and outputting the risk level to a risk control system. And the data management efficiency is improved.

The invention relates to the technical field of data privacy protection, in particular to a data privacy protection management system based on artificial intelligence, which is characterized in that a resource library building module is used for extracting privacy protection knowledge from multi-source data and building a structured resource library; and the privacy classification module classifies data privacy information and evaluates a sensitive level through intelligent identification and analysis, and provides a basis for a differential protection strategy. The privacy protection module fuses privacy types and sensitive levels, constructs a core policy engine, and comprises a privacy desensitization sub-engine, an access control sub-engine and a privacy risk prediction sub-engine, so as to realize automatic generation and dynamic adaptation of protection policies. And the privacy management module defines an execution sequence and a collaboration mechanism of each sub-engine according to a pre-configuration process, generates a privacy protection method covering the full life cycle of the data, ensures end-to-end collaboration management of links such as data acquisition, storage, use and sharing, and remarkably improves the compliance, efficiency and risk prevention and control capability of data privacy protection.

Disclosed in the present invention are a large language model federated fine-tuning method and apparatus based on gradient compression. The method comprises the following steps: constructing, on the basis of a gradient tensor generated during fine tuning of a large language model, a raw data set having a time series relationship, performing inference by means of an autoencoder to obtain a reconstructed gradient data set, and constructing a reconstruction loss function to optimize the autoencoder; and initializing a base model of the large language model as a global model at a server end, the server end updating the global model to a client, using a pre-trained encoder to obtain a compressed gradient at the client, and at the server end, using a pre-trained decoder to decode and aggregate the compressed gradient, and then updating the global model. The present invention can improve the fine-tuning efficiency of the large language model and reduce computing resource requirements while ensuring data privacy protection, and is suitable for application scenarios such as communication optimization improvement and privacy protection enhancement in the process of scientific computing-oriented large model fine-tuning and training.

The invention belongs to the technical field of multi-energy micro-grid cooperative regulation and control. The invention discloses a multi-energy micro-grid coordinated regulation and control system based on cross-layer knowledge injection and federated distillation. The system is characterized by comprising a cross-layer knowledge injection network; a federal knowledge distillation module; and the uncertainty map regulation and control module is used for constructing an MEMG uncertainty association map, learning the influence weight of each uncertainty factor on a regulation and control decision through a map attention mechanism, and realizing a dynamic risk avoidance strategy. The invention discloses a multi-energy micro-grid cooperative regulation and control method based on cross-layer knowledge injection and federated distillation. The method is characterized by comprising the following steps: step 1, MEMG topology knowledge coding; 2, cross-layer knowledge injection training is carried out; step 3, federal knowledge distillation optimization; and step 4, performing uncertainty map regulation and control. According to the system and the method, the precision, the robustness and the data privacy protection capability of MEMG regulation and control are improved, and the system and the method are suitable for efficient collaborative optimization of the park-level multi-energy microgrid.

The invention discloses a data privacy protection method for an industrial internet platform, and relates to the technical field of data security and privacy protection. According to the method, the sensitive information is accurately identified and deeply analyzed through the sensitive information feature library and the hierarchical matching algorithm, the problem of insufficient accuracy and flexibility during large-scale data processing is solved, the accuracy and reliability of data desensitization are remarkably improved, personal privacy is effectively protected, data availability is maximized, and the method is suitable for large-scale data processing. A hierarchical processing algorithm and a self-adaptive desensitization rule base are utilized, desensitization rules are dynamically adjusted according to dynamic access requirements and sensitivity levels of data, data security and availability are balanced, accurate protection under different scenes is ensured, system adaptability and flexibility are improved, and the data access process is monitored in real time, anomaly detection and rule verification are performed, so that the data access efficiency is improved. And the desensitization rule is dynamically adjusted, so that the security and reliability of the system are enhanced, the user credibility is improved, and the transparency and credibility of the data processing process are ensured.

The invention relates to the technical field of wires and cables, and particularly discloses a wire and cable test system based on sensor data acquisition, which comprises a data acquisition module, an intelligent analysis module and a function output module. The intelligent analysis module integrates technologies such as multi-modal signal blind source separation, mechanism-data hybrid drive deep learning, federated learning and digital twinborn linkage and self-supervised learning optimization, and forms a layer-by-layer progressive analysis system. Interference components are effectively eliminated through a signal separation technology, meanwhile, the adaptability of the model to different working conditions and different cable types is improved by means of fusion of mechanisms and data, model collaborative optimization under multi-site data privacy protection is realized through federated learning, and a self-supervised learning-driven digital twinborn evolution mechanism is combined, so that the multi-site data privacy protection model is optimized. The problems that a traditional analysis technology is insufficient in generalization ability, simulation precision is prone to drifting, and privacy risks exist in data sharing are solved, and the accuracy of fault recognition and the long-term adaptability of the system are remarkably improved.

The invention discloses an emergency rescue data privacy protection and collaborative analysis system based on federated learning, and belongs to the technical field of emergency rescue data privacy protection and collaborative analysis. A dynamic trust evaluation module; the federated learning privacy protection engine module is used for realizing cross-domain data collaborative training by adopting a secure multi-party computing model; and the block chain intelligent contract module is used for deploying a data access control strategy and a credible auditing rule. According to the method, a global trusted environment is established through an end-side cloud three-level trust chain, the reliability of equipment is quantified through dynamic trust evaluation, data collaboration of privacy protection is realized through federated learning, an auditing strategy is automatically executed through an intelligent contract, an efficient encryption technology for guaranteeing safety and effectively reducing resource consumption is adopted, and the urban emergency data safety problem is solved.

The invention discloses a block chain technology-based review data information security tamper-proofing method, and relates to the technical field of information security, the tamper-proofing method comprises the following steps: a data encryption storage step, a data evidence storage step, a trusted computing step and a data verification step, when a computing request is received, the trusted computing step is executed, and the trusted computing step is executed; whether the identity of a requester conforms to the access control strategy is verified through the smart contract, after verification is passed, the smart contract generates a calculation task event, the oracle monitors and captures the event, then a calculation task is transmitted to the trusted execution environment, ciphertext data is obtained in the trusted execution environment and decrypted, and the calculation task is sent to the target server. The method has the advantages that the balance of data privacy protection and integrity verification is realized through a hierarchical architecture combining on-chain evidence storage and off-chain encrypted storage, sensitive review data is stored in an off-chain database in an encrypted form, the security of the review data is improved, the security of the review data is improved, and the security of the review data is improved. Only data fingerprints are recorded on the block chain, data privacy is protected, and data integrity can be verified through the block chain.

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 discloses an integrated federated learning optimization method based on clustering weight sampling, and the method specifically comprises the following steps: a federated learning system comprises a plurality of clients and a server, and the server calculates the similarity between the clients through model updating information uploaded by the clients, clustering the clients by adopting a dynamic clustering method according to the similarity; the server carries out secondary clustering according to a set sampling rule and judges whether a first-stage iteration threshold value is reached, all the clients obtain a latest global model and freeze a model feature recognition layer for fine tuning, the server collects parameters of all the clients after fine tuning, and then the parameters are clustered according to similarity and are subjected to secondary clustering according to the sampling rule; and combining into an enhanced global model through an ensemble learning strategy. The method can be widely applied to data privacy protection scenes in the fields of medical image analysis, financial risk control, intelligent transportation and the like, and a new technical solution is provided for efficient application of federal learning in a heterogeneous environment.

The invention discloses a customer privacy protection method based on homomorphic encryption and data desensitization, and particularly relates to the technical field of data privacy protection. Original customer data are collected, data cleaning, standardization and sensitive information identification and classification are carried out, the identified customer sensitive information is dynamically calculated, and differential privacy noise parameters are applied, so that the customer privacy protection is realized. The method comprises the following steps: generating dynamic desensitization data, encrypting key numerical sensitive information in the generated dynamic desensitization data, generating ciphertext data, and constructing a security calculation and analysis platform which comprises a ciphertext calculation engine and a security query interface and is used for executing ciphertext addition and multiplication operations on the generated ciphertext data to form an encryption analysis result. Authorization decryption and result delivery are carried out on the encrypted analysis result, and the privacy protection effect and the data use condition of the whole process are recorded in an audit log, so that traceable privacy protection closed-loop management is realized.

The invention provides a medical data-oriented efficient privacy computing fusion engine method and system, and the method comprises the steps: constructing a medical data encryption and decryption system as a basis, and carrying out the standardized packaging of federated learning, homomorphic encryption, secure multi-party computing and other technologies, a unified technical framework comprising a data encryption interface, a ciphertext calculation interface and a result decryption interface is defined; an intelligent task analysis mechanism is adopted, high-level medical analysis requirements can be automatically decomposed into privacy calculation subtask sequences, and multi-technology collaboration is achieved through an execution engine; meanwhile, the functions of resource monitoring and performance analysis are achieved, technology switching and parameter adjustment and optimization during operation are supported, and a self-optimization execution environment is formed. And finally, displaying through a friendly application interface and visualization. According to the method, the problems of single technical route, low calculation efficiency, poor resource utilization rate, serious memory limitation and the like in medical data privacy protection are solved.

The invention discloses a vertical federal learning feature selection method based on context awareness, and belongs to the technical field of artificial intelligence and data privacy protection. According to the method, firstly, an unsupervised sparse network is utilized at a client to initialize the importance of local features so as to accelerate convergence and reduce calculation complexity; and then, obtaining the embedded representation of each client in a pre-training stage, and screening the embedded representation by combining context features through a server side, thereby indirectly identifying key features. In the feature selection stage, the client side performs local feature screening according to the significant embedded index issued by the server, and the influence of irrelevant features on calculation and communication is further reduced. According to the invention, an attention mechanism is introduced to dynamically evaluate contributions of different participants, so that fair weight distribution is realized. According to the method, through staged joint optimization, the communication and calculation cost in the federation training process is effectively reduced, and meanwhile, the prediction precision and interpretability of the model are improved.

The invention discloses a sensitive data protection-oriented generative model differential privacy leakage prevention method and system, and belongs to the technical field of artificial intelligence and data privacy protection. Constructing a forward diffusion and reverse sampling process based on sample information and a diffusion generation model framework; calculating the semantic sensitivity of the current generation state in each step of reverse sampling through a self-adaptive differential disturbance regulation algorithm, and regulating the noise intensity of the current step according to self-adaption; privacy expenditure brought by each round of disturbance is monitored through a verifiable privacy budget tracking mechanism, and real-time tracking privacy budget consumption is obtained; identifying high-similarity samples as high-risk areas in each round of generation by generating a risk perception feedback optimization strategy; and the model is guided to be far away from the privacy sensitive area by dynamically enhancing the disturbance intensity of high-similarity samples. According to the method, the privacy protection capability of the model in the sensitive field is effectively improved while high-quality data generation is realized.

The invention relates to the technical field of power system frequency modulation control, in particular to a thermal power generating unit collaborative frequency modulation method and system based on block chain federated learning. According to the method, a decentralized P2P network is constructed through a block chain technology, and model parameters are initialized based on a secret sharing mechanism; carrying out local training by adopting an adaptive federal near-end optimization algorithm, measuring data isomerism through KL divergence and dynamically adjusting a regular coefficient; carrying out noise addition on a sample by adopting differential privacy, and carrying out homomorphic encryption to realize parameter encrypted transmission; selecting representative nodes based on a block chain consensus mechanism to execute parameter aggregation in the encryption domain; and multi-unit collaborative frequency modulation is realized through model prediction control. According to the method, the problems of data privacy protection, heterogeneous data adaptation and decentralized cooperative training are solved, and the stability of power grid frequency regulation and control is improved.

The invention discloses a federated learning model compression method based on data-free knowledge distillation and multi-teacher knowledge distillation, and belongs to the technical field of federated learning. According to the technical scheme, a central server is used for pre-training a model to constrain random noise and labels, a loss function is optimized after the model is input into an image generator, and intermediate features are extracted and stored in a feature pool; and after the client uploads the model, sampling from the feature pool to generate a diversified composite image as training data, and training a local aggregation model by combining a pre-training model and a client integrated model as teachers. The beneficial effects are that the FedPET method provided by the invention combines data-free knowledge distillation and integrated knowledge distillation, and innovatively introduces channel feature exchange and multi-teacher model guidance, so that the model compression efficiency, generation diversity, data privacy protection, communication efficiency and accuracy of federal learning are significantly improved; the method has obvious advantages especially in heterogeneous data distribution and resource limited scenes, and an efficient, safe and accurate solution is provided for practical application of federated learning.

The invention discloses a channel security transmission method for protecting data privacy of an edge gateway of the Internet of Things, and relates to the technical field of security and privacy protection of the Internet of Things, and the method comprises the steps: constructing an edge gateway security architecture, generating a quantum key through a QKD module based on the constructed edge gateway security architecture, and generating sensitivity data through an IPS vNSF; calculating noise based on the sensitivity data, generating disturbance data based on the noise, calculating a shared key, generating an encryption key in combination with the quantum key and the shared key, segmenting the encryption key through Shamir secret sharing, and encrypting and decrypting the key; encrypting the noise by using the decrypted key, generating global noise based on the encrypted noise, and calculating secondary disturbance data based on the global noise; and generating coded data based on the secondary disturbance data, and generating reconstruction data based on the coded data. According to the invention, the comprehensive security authority of the edge gateway in the aspects of key security and privacy protection intensity is improved.

The invention discloses a blockchain zero-knowledge proof collaborative credit risk control data privacy protection query method, and belongs to the technical field of financial science and technology and information security. According to the method, a four-layer architecture of a feature importance analysis layer, a zero-knowledge proof verification layer, a block chain intelligent contract layer and an audit log layer is constructed, so that the contradiction between privacy protection and data availability in a traditional credit risk control system is solved. According to the system, an SHAP value is adopted to calculate feature importance, a zk-SNARKs verification circuit is deployed for high-sensitivity features, zk-STARKs lightweight verification is adopted for low-sensitivity features, and differential privacy protection is achieved. And dynamically adjusting the complexity of a verification algorithm according to the credit risk score of the user, and realizing dynamic authority control based on a verification result through an intelligent contract. The verification metadata is recorded to the block chain to form a tamper-resistant auditing log, and it is ensured that the query process is transparent and traceable. According to the method, the real-time risk control requirement is met while the privacy of the user is protected, and a compliant and efficient data query solution is provided for financial institutions.

The invention discloses a method and a system for realizing remote operation and maintenance of equipment through cooperation of an industrial intelligent gateway and a cloud platform, and relates to the technical field of remote operation and maintenance of industrial Internet of Things, and the method comprises the following steps: the gateway collects multi-modal data of the equipment and locally preprocesses the multi-modal data to generate a structured data packet; lightweight Transform model forward propagation is executed, and an encrypted feature vector is extracted and uploaded to a cloud platform; the cloud platform adopts federated learning to update model parameters based on the feature vectors, issues the model parameters to the gateway, and synchronously constructs equipment digital twin bodies; analyzing the natural language operation and maintenance instruction through the digital twin to generate a control command; an equipment association graph is constructed, a fault propagation path is analyzed by using a graph convolutional network (GCN), and an early warning is generated; and after executing the control instruction, the gateway feeds back state data to the cloud platform, updates the digital twin and model input, and forms closed-loop control. According to the invention, collaborative optimization of data privacy protection, cross-device fault prediction and intelligent decision is realized, and the operation and maintenance efficiency and security are improved.

The invention discloses a multi-modal training data desensitization and traceability management method, which comprises the following steps of: firstly, carrying out structured preprocessing on original multi-modal data, and then generating text weighted image features, image weighted text features and a key cross-modal attention graph through cross-modal feature extraction and collaborative attention fusion; on the basis of the attention graph and a privacy strategy library, entity identity priori detection is carried out innovatively in combination with a knowledge graph, and self-adaptive collaborative prediction is carried out on sensitivity by utilizing Bayesian inference; and finally, a refined desensitization plan is generated through variational optimization, and accurate desensitization of the original data is realized. By means of the mode, the defects of inconsistent desensitization and one-step desensitization in the aspect of multi-modal data privacy protection in the prior art are overcome, it is ensured that sensitive information is fully protected, meanwhile, the inherent value and training availability of data are reserved to the maximum extent, and the data privacy protection efficiency is improved. And a high-quality, compliant and valuable training data set is provided for the artificial intelligence large model.

The invention discloses a data privacy protection method based on multi-party security computing and a block chain, and the method comprises the following steps: 1, creating a multi-party security computing task request, and distributing the multi-party security computing task request to each participant node; 2, performing Kronecker perturbation coding on local original data, and performing Hash processing through an SHA-256 Hash algorithm to generate a perturbation Hash value; 3, recording access behaviors, and generating an access behavior track sequence; step 4, calculating a Soft-DTW distance between the access behavior track sequence and the historical access track sequence; 5, constructing an adjacent matrix and a Laplacian matrix, and carrying out eigenvalue decomposition; step 6, calculating the Euclidean distance between the access embedding vector and the legal access embedding vector; and 7, generating a result abstract and writing the result abstract into a task recording module. According to the method, the Kronecker perturbation coding and the Hash algorithm are combined, so that the data privacy protection, the calculation transparency and the compliance verification capability are improved.

The invention relates to the technical field of power grid data privacy protection, in particular to a smart power grid data privacy protection system and method based on a homomorphic encryption technology. The system protects data privacy of an intelligent power grid based on a homomorphic encryption technology, and comprises eight modules: a data acquisition module is connected with an intelligent electric meter in a multi-protocol manner in a 15-minute period to adopt electric data; the data filtering and preprocessing module optimizes data quality; the homomorphic encryption module generates a ciphertext capable of being added and multiplied by using an FHE algorithm; the multi-level key management module manages multiple types of keys in the whole life cycle; the secure transmission module transmits the ciphertext by means of a hybrid protocol; the data analysis module decrypts and analyzes the ciphertext in the data center; the decryption module decrypts after authorization; and the auditing and monitoring module records operation and monitors abnormity. According to the invention, it is ensured that the power use data is kept in an encrypted state in the whole life cycle of collection, transmission and analysis, the data is available and invisible, and the data privacy protection level and quantum attack resistance of the smart grid are improved.

The invention relates to the technical field of traffic engineering facility monitoring, and particularly discloses a traffic engineering facility safety state monitoring system and method based on the Internet of Things, and the method comprises the steps: deploying distributed intelligent sensing nodes comprising multiple types of sensors and edge calculation units at key monitoring points of facilities; nodes collect data by adopting a dynamic self-adaptive sampling mechanism, and an edge computing unit preprocesses the data and performs preliminary abnormity early warning; the data is transmitted to a cloud platform through a hierarchical communication network; the cloud platform performs deep analysis by using a federated learning safety evaluation model, and calculates a safety index to generate an evaluation report; and when the monitoring value exceeds the threshold value, four-stage early warning is triggered corresponding to different inspection or disposal processes. According to the invention, through the whole process optimization of dynamic sampling-layered transmission-federated learning evaluation-multi-stage early warning, efficient, accurate and low-consumption monitoring of the traffic engineering facility safety state is realized, and the contradiction between data privacy protection and evaluation precision is solved at the same time.

The invention relates to the technical field of data privacy protection and artificial intelligence crossing, in particular to a differential privacy and contrast learning fused data synthesis method, which comprises the following steps: S1, data acquisition and clustering: acquiring a real training data set, and clustering the data set by using a DBSCAN algorithm; s2, establishing double discriminators: designing a double discriminator framework comprising a privacy discriminator and a utility discriminator, injecting adaptive differential privacy noise based on a clustering structure into gradient updating by the privacy discriminator to realize privacy protection, and focusing on keeping the quality and authenticity of generated data by the utility discriminator. Precise balance between privacy protection and data utility is realized through a double-discriminator architecture. The privacy discriminator focuses on differential privacy constraints to ensure that the generated data meet strict privacy requirements; the utility discriminator restrains the data distribution consistency through the Wasserstein distance, effectively reduces the damage of the noise to the data utility, and solves the problem that the privacy and the utility are difficult to consider in the prior art.

The invention discloses an Internet of Things video monitoring big data privacy protection and efficient retrieval method based on artificial intelligence, belongs to the crossing field of artificial intelligence, Internet of Things and information security, and is suitable for vehicle-mounted, park, battery swap stations and other scenes. The method comprises the following steps: firstly, establishing a self-adaptive acquisition framework, and realizing multi-protocol switching and video preprocessing; secondly, extracting privacy information through an improved YOLO algorithm and a Graph-Cut technology, and combining reversible watermark embedding; constructing a multi-dimensional privacy level model for hierarchical encryption, and matching two-level storage and three-level index; then, two-factor authentication authorization is performed to extract privacy and optimize retrieval; and finally, the system state is monitored in real time and self-adaptive optimization is performed. According to the method, the problems of protocol heterogeneity, insufficient privacy protection, low retrieval efficiency and the like can be solved, and privacy security, storage overhead and retrieval efficiency are balanced.

The invention provides a federal learning optimization method and system for multi-source heterogeneous cardiovascular disease risk prediction, and relates to the technical field of data processing, and the method comprises the steps: processing original cardiovascular disease data, and obtaining standardized feature data; based on the standardized feature data, a differential training algorithm corresponding to the data structure of the standardized feature data is adopted for local calculation, and updated encrypted local parameters are generated; the updated encrypted local parameters are transmitted to a parameter server in a federated learning framework, the parameter server adopts a geometric mean algorithm based on secure multi-party calculation to perform secure aggregation and geometric mean integration on local parameter updates from different sources in an encryption state, and a preliminary global parameter set is generated. According to the method, data privacy protection is enhanced through a federated learning framework, the multi-source heterogeneous data fusion efficiency is improved, and parameter integration is optimized to enhance model performance.

The invention discloses a video image generation method and system driven by edge computing, and relates to the technical field of computer vision and edge computing, and the method comprises the steps: a terminal device initiates a generation task and perceives resources; dynamically dividing the generated task into a plurality of sub-tasks through a decision model based on the perception information, and deciding the execution position of each sub-task at the terminal, the edge node or the cloud; the nodes cooperatively execute the distributed subtasks and exchange intermediate data; and a final synthesis result is fed back to the terminal. The system comprises a terminal module, an edge node module, a cloud data center module and a collaborative management module. According to the invention, through terminal-edge-cloud cooperative computing, delay of a generation task and cloud bandwidth consumption are effectively reduced, data privacy protection is enhanced, system resource utilization efficiency and expandability are improved, and the method is especially suitable for video image generation application scenes with high real-time performance requirements.

The invention discloses an interpretable differential privacy parameter adaptive method for a generative medical record, and relates to the technical field of medical data privacy protection and artificial intelligence. The method comprises the following steps: acquiring an original medical record data set, and identifying a field containing sensitive information; constructing a multi-dimensional evaluation model, and respectively calculating the sensitivity score of each field and the utility score of the downstream task; dynamically allocating differential privacy budget parameters for each field by using a constraint optimization algorithm based on a game relationship between sensitivity and utility; in the training process of the generative model, corresponding noise is injected into a gradient or an input layer according to the distributed parameters; and finally, generating a synthetic medical record and outputting an interpretability report of privacy parameter distribution. By automatically adjusting the DP parameters, the data features of the low-sensitivity and high-value fields are reserved while the high-sensitivity fields are ensured to obtain strong privacy protection, the problem that privacy protection and data availability are difficult to consider in medical data sharing is effectively solved, and a transparent auditing basis is provided.

The invention provides a clinical data management system, a clinical data privacy protection method, equipment and a medium. Obtaining an identity identifier corresponding to the medical institution identity information of the requester and a patient anonymous identifier associated with the medical record text data; performing medical semantic analysis on clinical contents in the medical record text data, identifying a plurality of sensitive entities containing diagnosis and treatment features, physiological indexes and personal identifiers of patients, associating medical contexts of the sensitive entities, and generating a specified replacement seed value for each sensitive entity based on anonymous identifiers of the patients; and for each sensitive entity, determining corresponding replacement content from a preset desensitization rule base according to the medical context of the sensitive entity and the replacement seed value. And finally, replacing each sensitive entity with the corresponding replacement content, and generating a desensitized clinical text associated with the anonymous identifier of the patient. According to the scheme, context-aware privacy protection can be realized in a clinical data sharing environment.

The invention discloses a privacy enhanced CPPS anomaly detection method based on longitudinal federated learning, and relates to the field of CPPS anomaly detection. According to the method, through longitudinal federated learning, an SCINet model and a Transform model are deployed at clients of a physical side and an information side respectively and are used for local deep feature extraction; in the feature uploading stage, feature compression processing is performed on the physical side data deep features and the information side data deep features, so that effective compression of the uploaded features is realized, and meanwhile, the sensitive information amount possibly leaked in the middle features is greatly reduced. In addition, a bidirectional collaborative optimization mechanism between the client and the server is constructed, and a local feature extraction strategy can be optimized in real time. According to the method, the data privacy protection effect is remarkably improved while the anomaly detection accuracy is guaranteed, the defects of a traditional anomaly detection method in the aspects of privacy security, feature compression effectiveness, bilateral collaborative optimization and the like are overcome, and the practicability and security of CPPS anomaly detection are effectively improved.