73 results about "Data anonymization" patented technology

Systems and methods for large language model (LLM)-agnostic data anonymization. Data anonymization includes data obfuscation (and data deobfuscation) to protect confidential information a user is going to send to an LLM service or application programming interface (API). Encryption can be used for data obfuscation and particularly, for securing data from unauthorized access. Likewise, decryption can be used for data de-obfuscation.

Identity nodes within bidirectional dual node hashchain (BDNH) bundles generate hash keys by processing partial metadata and linking them through identity graphs. Translation nodes perform identity resolution, data anonymization, and re-keying by transforming original identifiers into anonymous IDs and re-key IDs. Laplace noise enabler capsules introduce Laplace noise into translation layer logs for differential privacy. BDNHs share metadata between adjacent bundles for secure and efficient data transmission. A cognitive analytics layer shared among all bundles verifies hashed encrypted data packets using a gossip protocol for real-time validation. An analytics workspace layer activates re-key IDs and creates digital tags for datasets using BERT transformers, linking data through knowledge graphs. The system manages parallel processing and includes machine learning modules, homomorphic encryption, federated learning, zero-knowledge proofs, attribute-based encryption, and anomaly detection. Bundles' AI engines work with the Laplace noise enablers to analyze data, optimize noise, and enhance the data clean room functionality.

Disclosed herein are methods for selecting and preparing patient data to facilitate the adoption and customized training of machine-learning models in clinical settings, particularly for radiation therapy treatment planning. The disclosed embodiments streamline the customized training process through an automated workflow that includes prefiltering patient metadata, retrieving relevant DICOM files, optional data anonymization, and generation of training data. The data is then organized into a format suitable for machine-learning training. The embodiments discussed herein reduce manual labor, minimize errors, and accelerate the integration of machine-learning into clinical workflows, enabling clinics to train and implement predictive models that replicate specific clinical practices, thereby enhancing treatment precision and improving patient outcomes.

This invention relates to the field of cross-border data security transmission, providing a method and system for cross-border VPN data anonymization transmission adapted to multi-regional compliance. The method includes: parsing the header of the original cross-border data packet to generate data flow context metadata; loading a set of compliance constraints and a task utility objective function accordingly; scanning the data payload in parallel to generate a pre-analysis feature set; constructing decision context information; instantiating the objective function and transformation strategy space; solving for the optimal transformation strategy through constraint optimization; compiling it into an optimal anonymization execution plan; executing the optimal anonymization execution plan to generate a new data feature set; combining it with non-sensitive data to reorganize and inject a new data payload; outputting a second cross-border data packet; encapsulating it into a third cross-border data packet and generating a decision audit log; comparing the compliance decision and audit label with the target end's local compliance policy to trigger a policy update. This invention integrates multi-regional compliance constraints and business utility objectives to construct an intelligent anonymization and audit closed-loop mechanism for cross-border data.

The present disclosure relates to data anonymization methods, apparatuses, and storage systems. In one example method, a storage system obtains a data access request, obtains target data based on the data access request, and performs anonymization processing on the target data to obtain anonymized target data. Then, the storage system sends the anonymized target data.

The invention discloses a data anonymization processing method and system based on differential privacy, and relates to the field of data anonymization processing, and the method comprises the steps: receiving a data retrieval request of a data user; retrieving from an original database to obtain an initial data set, and constructing a differential privacy parameter search space; randomly determining a first differential privacy parameter in the differential privacy parameter search space, and performing data anonymization processing on the initial data set; performing data privacy evaluation and data availability evaluation on the first to-be-evaluated data set to obtain a first data privacy degree and a first data availability degree, and generating a first parameter fitness; performing parameter optimization based on the first parameter fitness and the first differential privacy parameter, and determining an optimal differential privacy parameter; and performing data anonymization processing on the initial data set to obtain a target data set, and feeding back the target data set to a data user. The problems that in the prior art, data privacy leakage cannot be quantized, reliable privacy guarantee is difficult to provide, and dynamic adaptability of differential privacy parameters is poor are solved.

This invention discloses a knowledge-based intelligent interpretation system and method for laboratory reports, comprising four sequentially linked modules: a data input module for receiving laboratory data transmitted from the LIS system, the data including patient laboratory indicators and basic patient information, supporting data anonymization, and the basic patient information including at least age, gender, pregnancy status, medical history, and family medical history. This invention effectively reduces patient consultations, allowing patients to understand their reports immediately, reducing anxiety, significantly improving patient satisfaction, avoiding inconsistent interpretations among doctors, enhancing the ability of general practitioners in community hospitals to interpret laboratory reports, and reducing unnecessary duplicate examinations through accurate interpretation.

A data anonymization pipeline system for managing holding and pooling data is disclosed. The data anonymization pipeline system transforms personal data at a source and then stores the transformed data in a safe environment. Furthermore, a re-identification risk assessment is performed before providing access to a user to fetch the de-identified data for secondary purposes.

The application discloses a kind of campus health management system and method based on group benchmark analysis, comprising: data acquisition module, for collecting the health and behavior data of student individual, and implementing data anonymization processing;Data processing and analysis module, including group clustering unit, for clustering the anonymized student individual data into homogenization analysis group according to student group attribute;Benchmark generation unit, for calculating the dynamic statistical benchmark band of one or more health indicators in each analysis group based on historical data within the set time window;Individual comparative analysis unit, for comparing each student individual data with the dynamic statistical benchmark band of corresponding health indicators of its analysis group, and calculating the relative deviation of individual health indicators;Early warning and report generation module, for triggering hierarchical early warning according to individual comparative analysis result, and generating visual health report. It can be widely applied to the field of campus health management technology.

The invention relates to the technical field of data security, and discloses a data desensitization method and system. The method comprises the following steps: analyzing a data access request of a user, and determining a target data structure; based on the target data structure, a target classification and grading label is determined, and the classification and grading label is generated by analyzing the initial service information and the system data source by using a large model in advance and is matched with the corresponding data structure in advance based on a preset matching rule; a target desensitization strategy is determined based on the target classification and grading label, and the desensitization strategy is determined by analyzing the classification and grading label of the corresponding data in the system data source through a large model in advance; obtaining to-be-accessed data based on the data access request; and performing desensitization processing on the to-be-accessed data by using the target desensitization strategy to obtain desensitized data, and sending the desensitized data to the user. According to the technical scheme, the data classification and grading efficiency and the desensitization strategy making accuracy can be improved.

The invention provides a data desensitization method based on bastion host operation and maintenance, a computer program product, electronic equipment and a storage medium, and the method comprises the steps: obtaining communication traffic of bastion host operation and maintenance; analyzing the communication flow to obtain standardized format data; inputting the standardized format data into a pre-constructed large model for sensitive data identification to obtain sensitive data and a corresponding sensitive degree; querying a pre-constructed permission desensitization strategy mapping table according to the sensitive data and the corresponding sensitive degree to obtain a desensitization strategy; and desensitizing the sensitive data according to the desensitization strategy. By implementing the application, accurate classification and identification of sensitive data can be realized, different desensitization strategies can be carried out according to different protocol types, various fine-grained desensitization can be realized, desensitization can be carried out according to operation and maintenance rights, and the desensitization process is more flexible and controllable.

Embodiments of the present disclosure relate to customized machine learning training of radiotherapy clinics. Disclosed herein are methods for selecting and preparing patient data to facilitate the employment and custom training of machine learning models in a clinical environment, particularly for radiotherapy treatment planning. The disclosed embodiments simplify the customized training process through an automated workflow that includes pre-filtering patient metadata, retrieving related DICOM files, optional data anonymization, and generating training data. And organizing the data into a format suitable for machine learning training. The embodiments discussed herein reduce manual labor, minimize errors, and accelerate the integration of machine learning into the clinical workflow, enabling clinics to train and implement predictive models replicating specific clinical practices, thereby improving treatment accuracy and improving patient outcomes.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for anonymizing unstructured data. In some implementations, a server can receive unstructured data. The server can automatically detect attributes in the unstructured data using a trained machine-learning model and can determine an amount of undetected attributes and detected attributes in the unstructured data. The server can simulate additional attributes for the unstructured data according to the amount of undetected attributes. The server can analyze a risk of disclosure in the unstructured data using the detected attributes and the simulated additional attributes. The server can modify the detected attributes according to the analyzed risk of disclosure and replace the detected attributes with the modified detected attributes in the unstructured data.

This invention relates to the field of data anonymization technology and provides a dynamic data anonymization method for electronic data, comprising the following steps: Calculating a query intent convergence score based on a received electronic data access request. In this invention, upon receiving an electronic data access request, the field range, filtering conditions, aggregation behavior, and join depth of the query statement are first analyzed, and a convergence score is calculated. The score directly maps to the granularity of field anonymization, ensuring that the degree of anonymization closely matches the query intent and avoiding excessive masking that could distort the results. Subsequently, external connectivity paths are traced based on quasi-identifiers and asset graphs. Information arbitrage risk is quantified according to multiplicative probability and sensitive value. If the risk exceeds the threshold, a cutoff command is immediately triggered, and high-risk fields are immediately isolated to prevent cross-domain assembly.

The invention discloses a privacy protection-oriented anonymized data watermarking method and device and a medium, and relates to the field of privacy protection, and the method comprises the steps: encrypting to-be-embedded original watermark information, and generating encrypted watermark information; performing error correction coding on the encrypted watermark information to obtain a watermark coding sequence, and converting the watermark coding sequence into a Gray code sequence; sorting the data records in the original anonymized data set based on preset attributes to obtain a plurality of record groups; based on the segmentation numerical value of the Gray code sequence, determining the number of records needing to be modified in each record group; and randomly selecting the data records corresponding to the number, and modifying the selected data records to generate the anonymized data containing the watermark. By combining cryptographic encryption, error correction coding and a data modification mechanism keeping anonymization, reliable traceability of illegal distributors is realized on the premise of ensuring that the data anonymization level is not reduced, so that an effective balance mechanism is established between data sharing and privacy security.

This application relates to the field of data anonymization technology, specifically to a method and system for anonymizing sensitive information in enterprise information databases. The method includes: extracting sensitive fields from the table structure of an enterprise's information database; calculating the first and second correlation degrees between any two sensitive fields to determine their correlation evaluation value; dividing all sensitive fields into multiple sets of related fields, obtaining the anonymization strength of each set, determining the anonymization level of each set, and selecting and executing the corresponding collaborative anonymization strategy. This application can completely block combined inference attack paths and ensure logical consistency of data during the anonymization process, achieving integrated collaboration between security protection and data utility.

This disclosure provides a data anonymization method, apparatus, electronic device, and computer-readable storage medium, which can be applied to the fields of data processing technology and finance. The data anonymization method includes: identifying data to be anonymized in a database and database components associated with the data to be anonymized; decoupling the database components from the data to be anonymized to obtain decoupled data to be anonymized; performing modified anonymization on the decoupled data to obtain anonymized data; and coupling the anonymized data and the database components to complete the data anonymization.

A system accesses database tables comprising sensitive data, collecting data elements corresponding to the sensitive data, reduces the data elements to a distinct list of data elements, generates a first script to generate map tables for each data element, wherein each map table comprises a first column to hold original values of the data element and a second column to hold anonymized values for the original values, generates a second script to scan the database tables to collect original values for each data element and populate the original values in a respective map table, generates a third script to anonymize the collected original values for each data element and populate the anonymized values in a respective map table, and generates a fourth script to update the original values using the corresponding anonymized values in the database tables based on the map tables.

The application provides an edge computing-based medical data anonymous distribution system, method, medium and terminal, receives collected initial medical data, processes the initial medical data into corresponding hierarchical anonymous medical data according to a medical data calling request received from a data consumption layer through an edge node, and displays the corresponding hierarchical anonymous medical data to a user. The application overcomes the technical problems of single medical data anonymization processing means, high risk of leakage and lack of correlation between anonymous data in the prior art, can balance privacy compliance and data correlation at the source of multi-source medical data, and is suitable for diversified medical data protection and distribution requirements.

A device, method, and system for treating data from legacy infrastructure is disclosed. Illustratively, the device memory stores computer executable instructions that when executed by the processor cause the processor to provide a dataset comprising a plurality of characters and provide a token table for tokenizing datasets. The token table includes mappings that define replacement tokens for characters in datasets. The instructions cause the processor to generate a tokenized dataset based on the dataset by (1) for each contiguous sequence of letter characters of the plurality of characters, determining a respective letter token having the same length as the respective contiguous sequence, and (2) generate the tokenized dataset by replacing each contiguous sequence of letter characters of the plurality of characters with the determined respective letter token.

Example embodiments of the present disclosure relate to machine learning (ML) model data privacy enhancements. In one aspect, a network entity trains a first machine learning ML model and notifies a second network entity to update information associated with the first ML model based on receiving a request to train the first ML model. The network entity receives a retraining request for the first ML model from the second network entity, the retraining request based on the updated information and indicating a change in privacy or agreement. The network entity retrains the first ML model based on the retraining request to generate a second ML model. In this manner, data privacy modifications and data anonymization enhancements are implemented, and data point changes are recorded.

This invention discloses a testing method, apparatus, device, medium, and product for isomorphic migration. The method includes: acquiring raw traffic information recorded by a browser plugin based on user operations on a source environment within a browser webpage, and a first response result from the source environment to the raw traffic information; converting the raw traffic information to obtain target traffic information; sending the target traffic information to a target environment to obtain a second response result from the target environment; and determining the test result of this isomorphic migration based on the first and second response results. This invention can solve the problem of traffic unavailability caused by data anonymization, thus improving testing efficiency.

The invention provides a data anonymization method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining a to-be-anonymized data file; based on an identifier identification method, obtaining an identifier identification result of the to-be-anonymized data file; calculating a data attribute identification degree score of each to-be-anonymized data column, and determining each identifier capable of realizing an anonymization requirement; determining each identifier to be processed according to a result of measuring the information amount of each identifier capable of realizing the anonymization requirement; and according to a preset constraint condition, processing the to-be-processed identifier until the constraint condition is met. By measuring the influence of anonymization processing on the inherent information of the to-be-anonymized data, each to-be-processed identifier and specific processing operation are determined in a targeted manner, double targets of balancing anonymization and retaining information are considered, and the inherent information of the to-be-anonymized data is retained to the greatest extent on the premise of meeting the anonymization targets.

This invention provides a multi-source fusion-based system and method for calculating peak fatigue levels in runners of up to 10,000 users. The system includes a data acquisition module, a data preprocessing and fusion module, a peak fatigue calculation module, a safety early warning and coordinated intervention module, a data security and privacy protection module, and a backup data acquisition unit. The core functionality involves collecting five-dimensional, multi-source, safety-related data on runners' physiological, exercise, environmental, individual, and subjective factors through the data acquisition module. A simplified version of non-invasive electromyography (EMG) signals is introduced to identify latent muscle fatigue. A dual-layer fusion architecture of "edge + cloud" is adopted, combined with a federated learning model, to achieve deep fusion of multi-source data while protecting runner privacy, constructing a personalized dynamic fatigue threshold model. The system calculates peak fatigue levels based on a real-time safety-oriented fatigue index, implements closed-loop intervention through a four-level graded early warning mechanism linking multiple terminals, and ensures data security through three-link redundant transmission, data anonymization and encryption. The backup acquisition unit ensures full coverage of data from up to 10,000 users.

A system and method for mitigating data loss in data anonymization processes. The method includes receiving, by a processing device, a first data item associated with first metadata, determining whether the first data item satisfies a sensitivity criterion, responsive to determining the first data item satisfies the sensitivity criterion, identifying, among a plurality of reference data items, a second data item that is closest to the first data item, wherein the second data item is associated with second metadata, determining whether a first similarity score the second data item satisfies a similarity criterion, responsive to determining the first similarity score satisfies the similarity criterion, generating synthetic data comprising the second data item and the first metadata, and using the synthetic data in training data for training an AI model to identify one or more patterns in the training data.

The embodiment of the invention discloses an anti-AI training medical data anonymization method and device, a medium and a program. The method comprises the steps of generating a fingerprint key stream based on an identifier of a data supplier and an identifier of a model research and development implementation party; based on the fingerprint key stream, embedding fingerprint information provided by a data supplier into medical data of the data supplier to generate a corresponding fingerprint-containing sample; obtaining and storing a unique hash value containing a fingerprint sample; sending the fingerprint-containing sample to a model research and development implementation party; when it is detected that the leaked sample exists, whether the leaked sample contains fingerprint information of a data supplier or not is detected; and if yes, determining a leakage source from the model research and development implementation parties which have sent the fingerprint-containing sample by detecting the unique hash value of the leaked sample, and generating a leakage traceability report and feeding back the leakage traceability report to the data supplier. According to the method, on the premise that the medical data diagnosis value and the AI model training precision are not affected, accurate tracing of medical data leakage is achieved, and a legal evidence storage basis is provided.

This invention specifically relates to a financing service management system based on multi-platform data sharing, encompassing the fields of data processing and financing service technology. It includes: a multi-source data acquisition and fusion governance module; an intelligent financing application and processing module; a financial institution collaboration and direct connection module; an intelligent risk control and data-driven decision-making module; and a financing dashboard and decision support module. In this invention, standardized interfaces and a federated learning framework enable secure acquisition and sharing of multi-source data. After data cleaning and standardization, the quality and usability are significantly improved, providing reliable data support for financing services. Simultaneously, through national cryptographic algorithms for data anonymization and compliance control, data acquisition and use are ensured to comply with relevant laws and regulations, reducing compliance risks.

This invention discloses a blockchain-based trusted carbon emission data sharing system, comprising: a data on-chain module for constructing a standardized data structure and recording data hash values ​​and on-chain index information through an on-chain contract; a trusted identification module for constructing multi-period carbon emission change trajectories and generating a set of deviation evolution indicators; an interpretation and interaction module for sending interpretation requests to carbon emission source nodes through a behavior request contract and collecting behavior interpretation packages submitted by carbon emission sources; a behavior judgment module for generating behavior confidence scores and spoofing risk markers through a behavior labeling contract; and a sharing control module for generating a trustworthiness score vector for carbon emission data through a trusted scoring contract and determining whether to trigger data anonymization processing or refuse sharing instructions. This invention achieves trusted identification, behavior verification, and dynamic access control in the multi-party sharing process of carbon emission data, improving data trustworthiness and sharing security.

The invention provides a data anonymization method and device, electronic equipment and a computer storage medium, and the method comprises the steps: carrying out the feature extraction of original data, and obtaining an original semantic vector set; the method comprises the following steps: distributing a differentiated dimension budget to each original semantic vector in an original semantic vector set, generating an anchor point constraint relationship of the original semantic vectors according to the original semantic vectors and an anchor point set, and finally processing the original semantic vectors in combination with the dimension budget corresponding to the original semantic vectors and the anchor point constraint relationship. The anonymous representation of the original semantic vector is generated, so that data privacy protection and analysis availability are considered while the anonymous representation is generated.