43 results about "Privacy Level" patented technology

The invention discloses a method and system for privacy data protection in a big data analysis process, and relates to the technical field of civil aviation data security and privacy protection, and the method comprises the steps: obtaining civil aviation multi-source heterogeneous data, extracting passenger identity features, travel track features and consumption behavior features, and generating a structured feature matrix; extracting an association relationship between fields through a convolutional neural network, calculating a field sensitivity initial value, generating a field sensitivity weight in combination with a field uniqueness index and the sensitivity initial value, and generating a privacy level mapping table containing the field sensitivity weight according to the sensitivity weight; calculating query sensitivity and a privacy budget parameter, generating a differential privacy noise parameter and a homomorphic encryption key, and generating a hierarchical protection strategy; performing noise injection and ciphertext conversion on the original data to generate a desensitized data set; passenger flow statistics is executed through secret state calculation, and a public analysis result is generated. Over protection of low-sensitivity data is avoided, and sufficient safety of high-sensitivity data is ensured.

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.

One embodiment of the present invention sets forth a technique for providing awareness of privacy-related activities. The technique includes determining a privacy level associated with a user of an extended reality environment. The technique also includes presenting, using an internal display of a headset, one or more internal indicators identifying a location of a bystander located in a real-world environment, wherein a level of detail of each internal indicator is based on the privacy level associated with the user. The technique further includes presenting, using an external display, one or more external indicators that include a monitoring indicator representing being captured by the headset and presented to the user via the headset, and further include a user activity indicator representing one or more activities of the user, wherein a level of detail of the user activity indicator is based on the privacy level associated with the user.

The invention relates to a method for realizing multi-modal large-model fine-grained privacy grading protection in an edge-cloud collaborative inference system, belongs to the technical field of large-model data security and privacy protection, and aims to solve the problems that privacy grading is rough, inference intermediate features are easy to leak and precision is damaged by protection measures in the prior art. According to the method, multi-modal data such as images and texts are received by edge computing nodes, deep semantic analysis and fine-grained segmentation are performed by using a local pre-trained multi-modal large model, and privacy scores are calculated to divide high and low privacy levels; uploading low-privacy data to a cloud computing center, and locally processing high-privacy data; and selectively shielding a middle layer feature map F generated by the edge based on the correlation degree and the contribution degree, then completing local reasoning, and uploading a result to be fused with the cloud. According to the method, precise fine-grained protection is achieved, the sensitive information leakage risk is remarkably reduced, and meanwhile the model reasoning precision and the system real-time performance are kept to the maximum degree.

The embodiment of the invention discloses a safety calculation method and device for intelligent workshop device coordination and a medium, and relates to the technical field of safety calculation, and the method comprises the steps: carrying out the on-chain device authentication of a plurality of device terminals in a target intelligent workshop in advance, so as to obtain the device production data corresponding to the target device terminal passing the authentication, performing sensitivity grading on the equipment production data, and determining a corresponding dynamic privacy level; matching a corresponding data encryption strategy according to the dynamic privacy level, determining a plurality of data encryption fragments corresponding to the equipment production data based on the data encryption strategy, and storing the data encryption fragments to a preset distributed node; and obtaining a current collaborative scheduling task of the target intelligent workshop, performing security calculation on the predetermined associated data encryption fragment through a predetermined edge calculation node based on a mixed framework of a confusion circuit and arithmetic secret sharing, generating a collaborative scheduling instruction, and feeding back the collaborative scheduling instruction to the target equipment terminal, therefore, safe calculation of equipment cooperative scheduling is realized.

The present invention concerns a method for managing a privacy level in a wireless network comprising an access point, AP, station and at least one non-AP station, the method comprising by a first station, either a non-AP station or the AP station sending, to a second station a message comprising an information indicating a set of privacy enhancement, PE, parameters to be obfuscated, the set of PE parameters corresponding to a privacy level requested by the first station, the second station being the AP if the first station is a non-AP station, the second station being a non-AP station if the first station is the AP, and receiving from the second station a response indicating whether the second station accepts the requested privacy level.

The invention discloses a hierarchical privacy protection data sharing method and device based on classification transformation perturbation, and the method comprises the steps: a terminal user freely selects a privacy budget, carries out the standardized preprocessing of numeric privacy data, carries out the classification of the preprocessed privacy data, and carries out the classification perturbation under the corresponding privacy budget. And carrying out classified restoration operation to output privacy-protected data, and uploading the classified and restored data and the corresponding privacy level to a server for hierarchical storage. And finally, grading data sharing is performed on the data under the multiple privacy grades, and the data users of the corresponding credibility grades can only obtain the disturbance reduction data under the corresponding privacy budget for data analysis. According to the method, a new solution is provided for high-availability hierarchical privacy protection mean value estimation.

An image grading privacy protection method and device and a storage medium are applied to the technical field of image encryption sharing, and the method applied to a first device comprises the following steps: sequentially encrypting a target image according to image encryption keys respectively corresponding to a plurality of privacy grades to obtain an encrypted image; acquiring encrypted public and private keys generated based on LWE and corresponding to each user level; according to a preset encryption algorithm and an encryption public key corresponding to the first target user level, encrypting an encryption private key of a second target user level to obtain a public value, the second target user level being any user level except the user level with the highest priority, and the first target user level being one priority higher than the second target user level; according to a preset encryption algorithm and the encryption public key corresponding to each user level, encrypting the corresponding image encryption key to obtain a plurality of ciphertexts; and sending the encrypted image and all ciphertexts and public values to a second device. And the data security and quantum security of image data and key transmission are improved.

The invention discloses a public institution internal data sharing method based on multi-source data fusion, and relates to the technical field of data sharing. According to the method, four-quadrant accurate division is carried out on multi-source fusion three-medicine structured data on the basis of sensitive risk scores and utility influence tolerance, differential noise adding strategies are matched in a targeted mode, and three kinds of deviation verification of medical quality, medical insurance fund risks and medicine supply and demand results and noise iterative optimization are combined, so that the multi-source fusion three-medicine structured data is obtained. And finally, through block chain intelligent contract encryption sharing, the problems that a traditional differential privacy technology is rigid in noise intensity setting and is difficult to adapt to data of different privacy levels are solved, dynamic balance between privacy protection strength and data availability is realized, meanwhile, the security and credibility of a data sharing process are guaranteed, and the data sharing efficiency is improved. On the premise that the multi-source heterogeneous public institution internal data meets the privacy security requirement, the practical value of supporting core services such as medical quality evaluation, medical insurance supervision and drug supply and demand allocation is reserved to the maximum extent.

The invention provides a data delivery method, device, medium and system, and the method comprises the steps: determining a target privacy calculation protocol according to a data type and a privacy level before data delivery, or determining the target privacy calculation protocol according to the data type; binding a data use constraint, a privacy calculation parameter and a participant digital identity by adopting an intelligent contract; and according to the target privacy calculation protocol and the bound smart contract, constructing a data delivery application program, and executing the data delivery application program to complete data delivery, the data delivery application program being an application program matched with the data product type. The problem that in the prior art, a unified application program development standard is not formed, so that data delivery tools are fragmented, and compatibility is poor is solved.

The invention discloses a privacy protection method and system based on Internet of Things data encryption transmission, and the method comprises the following steps: a data collection step: collecting original data from at least one Internet of Things device, the Internet of Things device including but not limited to a sensor, an actuator, a smart home device or an industrial Internet of Things node, the original data comprises one or more of environment data, user behavior data and equipment state data, the acquisition frequency is dynamically adjusted according to equipment types and application scenes, the range is one time per second to one time per hour, and a tamper-proof mechanism is adopted in the acquisition process to ensure the data integrity; the core advantage of the method is that refined and quantitative evaluation of the data privacy level is realized by introducing an innovative privacy score calculation formula.

One embodiment of the present invention sets forth a technique for providing awareness of privacy-related activities. The technique includes determining a privacy level associated with a user of an extended reality environment. The technique also includes presenting, using an internal display of a headset, one or more internal indicators identifying a location of a bystander located in a real-world environment, wherein a level of detail of each internal indicator is based on the privacy level associated with the user. The technique further includes presenting, using an external display, one or more external indicators that include a monitoring indicator representing being captured by the headset and presented to the user via the headset, and further include a user activity indicator representing one or more activities of the user, wherein a level of detail of the user activity indicator is based on the privacy level associated with the user.

The invention discloses a multi-party marketing data privacy collaboration method based on a cross-chain technology, relates to the technical field of data security and privacy computing, and aims to solve the technical problems that a current centralized data collaboration scheme has trust dependence, security risks and efficiency bottlenecks and is difficult to meet cross-mechanism data security fusion and efficient collaboration requirements. The method is implemented on the basis of a multi-party marketing data privacy collaboration system based on the cross-chain technology, and the multi-party marketing data privacy collaboration system based on the cross-chain technology comprises a data provider node cluster: deploying a client with a feature engineering module and a homomorphic encryption engine, the cross-chain relay network is used for generating encrypted feature data and initiating cross-chain asset registration, the cross-chain relay network comprises a verification node pool and a cross-chain protocol gateway, and the verification node pool verifies a cross-chain transaction through a consensus mechanism. The method has the advantages of high credibility, improved privacy level and stable collaboration efficiency.

The embodiment of the invention provides a task scheduling method and device for itinerary planning, a storage medium and a program product. In the scheme, firstly, based on multi-modal demand semantics, a system is enabled to adaptively identify an internal structure of a travel task, global complexity, real-time performance and a privacy level are intelligently mapped to a subtask level, and a'global demand-subtask constraint-end-edge-cloud node scheduling 'parameterized control link is formed; secondly, the reinforcement learning optimizer can dynamically adjust the scheduling weight, and strategy self-adaptive evolution is achieved. Furthermore, an execution-monitoring-feedback-rescheduling closed-loop system can be constructed, intelligent degradation and context snapshot migration are triggered when the node state deviates, seamless connection and continuous execution of a task chain among cross-hierarchy heterogeneous nodes are ensured, and finally, quick response is realized on the premise of ensuring the accuracy of a recommendation result of the travel information, so that the recommendation efficiency of the travel information is improved. And the system performance and the user experience can be balanced.

One embodiment configures multiple parameters that can be used to generate artificial data from raw data, adjusting the privacy level in the artificial data. The embodiment fits a distribution type to variables in the raw data. The embodiment uses a desired privacy level and distribution type to adjust a noise level, where the noise level corresponds to the desired privacy level. The embodiment uses a distribution type and noise level to generate artificial data that achieves the desired privacy level by including noise data corresponding to the noise level.

The invention discloses a single-encryption multi-level differential privacy output method based on attribute driving, and the method comprises three entities: an attribute management mechanism, a data owner and a data visitor, and comprises the following steps: the attribute management mechanism inputs data information to a computer system, and executes system initialization based on security parameters; generating a system master key, a public parameter and directory metadata, and signing and publishing the system master key, the public parameter and the directory metadata; the data owner obtains directory metadata and the signature result for verification, and output results of a plurality of privacy levels are generated through incremental noise accumulation; an encrypted data packet is formed by combining attribute encryption and authentication encryption mechanisms; the data accessor obtains the encrypted data packet and the directory metadata, consistency verification is carried out, decryption operation is carried out according to the hierarchy from high to low after verification, and a highest available hierarchy result is output. According to the method, the privacy protection capability is improved while the data availability is ensured, and the method is suitable for a multi-level permission data sharing scene.

The invention provides an operation and maintenance method and device, equipment, a medium and a program product, relates to the technical field of communication, and is used for improving the resource utilization rate and operation and maintenance efficiency of operation and maintenance, and the specific technical scheme is as follows: obtaining operation and maintenance data and first state information of terminal equipment, the first state information comprising a terminal computing power load and a terminal network bandwidth; under the condition that it is detected that second state information of the cloud equipment meets a preset condition, an operation and maintenance allocation decision value is obtained through calculation based on the terminal computing power load, the terminal network bandwidth and feature information corresponding to the operation and maintenance data, and the second state information comprises the cloud computing power load and the cloud storage remaining capacity; the feature information at least comprises task complexity, data privacy and real-time requirements; based on the privacy level and the operation and maintenance allocation decision value corresponding to the operation and maintenance data, determining an operation and maintenance device for processing the operation and maintenance data, the operation and maintenance device being a terminal device or a cloud device; the method is applied to an operation and maintenance early warning scene.

In general, certain embodiments described herein relate to a method for performing inferences on noisy compressed data. The method includes receiving, from a client, the noisy compressed data at a machine learning system. The noisy compressed data is transmitted over a network with the machine learning system being a machine learning as a service configuration. An amount of noise in the noisy compressed data is based on a privacy level. The method further includes classifying, by a classifier of the machine learning system, the noisy compressed data by executing an inference on the noisy compressed data to obtain classification results, and transmitting the classification results to the client via a network. The inference is executed by a trained model, which is trained on noisy compressed training data. By transmitting noisy compressed data, the privacy of actual data is preserved.

This invention discloses a single-pass encryption multi-level differential privacy output method based on attribute-driven principles. The method comprises three entities: an attribute management organization, a data owner, and a data accessor. The steps include: the attribute management organization inputs data information into a computer system and performs system initialization based on the security parameters, generating a system master key, public parameters, and directory metadata, which are then signed and published; the data owner obtains the directory metadata and the signature result for verification, generating multiple privacy-level output results through incremental noise accumulation; and combines attribute encryption and authentication encryption mechanisms to form an encrypted data packet; the data accessor obtains the encrypted data packet and directory metadata, performs consistency verification, and upon successful verification, performs decryption operations from highest to lowest privacy level, outputting the highest-level result. This invention improves privacy protection capabilities while ensuring data availability and is suitable for multi-level access control data sharing scenarios.

The invention relates to the technical field of data security, and discloses a privacy data desensitization method and system based on a pre-training language model.The method comprises the steps that data normalization is conducted on original text data, and standardized text data is obtained; detecting sensitive entity fragments through a pre-training language model, and extracting boundary information to generate sensitive entity position information; extracting sensitive entity contents to obtain a to-be-desensitized entity set; constructing a context semantic association network, deducing a dynamic privacy level based on a topological relation to obtain a hierarchical privacy identifier, performing semantic adaptation with a predefined desensitization rule base to obtain a semantic calibrated strategy, and performing format-preserving transformation on sensitive entity content to obtain desensitized text data; verifying semantic coherence of the desensitized text data by using a pre-training language model, if the semantic coherence passes the verification, outputting a target desensitized text, and if the semantic coherence does not pass the verification, feeding back to reconstruct a context semantic association network; according to the invention, the privacy data desensitization efficiency can be improved.

The invention provides an industrial manufacturing scene distributed learning-oriented space-time fine-grained privacy protection method and system. The method comprises the following steps: constructing distributed learning of an industrial manufacturing scene; spatial hierarchical modeling and privacy level division are carried out; performing production line node side configuration and execution of differential privacy, homomorphic encryption and data desensitization; designing a time dynamic privacy regulation and control mechanism; optimizing and aggregating spatio-temporal joint privacy parameters; global aggregation and privacy budget recovery; and model updating and stage feedback adjustment. According to the method, the space security level difference and time stage change characteristics in the industrial manufacturing process are comprehensively considered, a hierarchical differential privacy, homomorphic encryption and data desensitization mechanism is introduced into the space dimension, and a dynamic privacy budget and encryption strength regulation and control mechanism is introduced into the time dimension, so that a space-time joint fine-grained privacy protection system is formed.

This invention provides a multi-level privacy protection method, device, and equipment for caregiving scenarios based on multimodal collaborative perception. By constructing a multimodal neural network model at the caregiving data acquisition terminal and fusing visual and auditory information for deep semantic analysis, it effectively overcomes the limitations of traditional single-modal perception in complex environments such as low light and occlusion. This invention employs a collaborative design of delay buffering and pre-judgment, forcibly placing privacy level determination before media stream distribution, fundamentally eliminating the risk of instantaneous leakage of privacy images due to inference delays. By introducing a two-level dynamic judgment strategy of prioritizing high-risk feature blocking and scene adaptive fusion, the robustness of the system in real-world home environments is improved. Based on the matching relationship between user permissions and privacy levels, this invention performs differentiated media stream processing on the terminal side, achieving fine-grained control of privacy protection granularity and perfectly balancing the dual needs of caregiving functions and privacy protection.

The invention discloses a context awareness type emotion AI agent management method and equipment, and mainly relates to the technical field of context awareness. The method and the device are used for solving the problems that a traditional architecture is difficult to meet high-concurrency processing requirements and complex emotional interaction requirements at the same time, the ability of intelligently awakening related memories according to situations lacks, and roles and states of intelligent agents are mixed in the prior art. Comprising the following steps: calculating a priority value of a corresponding memory node according to semantic similarity, a privacy level in the memory node, an emotion weight, emotion data and an emotion conflict coefficient of current emotion data; meanwhile, according to a preset weight updating formula, updating and calculating an edge weight between the preset topic node and the matched memory node; and according to the falling interval of the priority value, determining whether to actively mention the memory node corresponding to the priority value, or mention the memory node corresponding to the priority value when the user mention the preset related topic corresponding to the memory node, or not waking up the memory node.

Various examples described herein are directed to systems and methods for managing an interface between a user and a user computing device. The user computing device may determine that an audio sensor in communication with the user computing device indicates a first command in a user voice of the user, where the first command instructs the user computing device to perform a first task. The user computing device may determine that the audio sensor also indicates a first ambient voice different than the user voice and match the first ambient voice to a first known voice. The user computing device may determine that a second computing device associated with the first known voice is within a threshold distance of the user computing device and select a first privacy level for the first task based at least in part on the first known voice.

The application provides a kind of space-time fine-grained privacy protection method and system for industrial manufacturing scene distributed learning.The method comprises: industrial manufacturing scene distributed learning construction;Space hierarchical modeling and privacy level division;Differential privacy, homomorphic encryption and data desensitization line node side configuration and execution;Time dynamic privacy control mechanism design;Space-time joint privacy parameter optimization and aggregation;Global aggregation and privacy budget recovery;Model updating and stage feedback adjustment.The application comprehensively considers the spatial safety level difference and time phase change characteristics in the industrial manufacturing process, by introducing hierarchical differential privacy, homomorphic encryption and data desensitization mechanism in the spatial dimension, introducing dynamic privacy budget and encryption intensity control mechanism in the time dimension, forming a space-time joint fine-grained privacy protection system.

The invention relates to the technical field of artificial intelligence, and discloses a memory management method and system, electronic equipment and a storage medium. The memory management method comprises the following steps: analyzing dialogue content in real time to generate short-term memory and storing the short-term memory; dynamically updating the memory intensity of each short-term memory, and promoting the short-term memory of which the memory intensity reaches a long-term memory threshold value into long-term memory storage; performing privacy level marking on the short-term memory and the long-term memory; and sensing the current situation in real time, searching related memories from the database according to the current situation, performing privacy level filtering on the related memories, and generating and executing active services in combination with the current context and the filtered related memories. According to the embodiment of the invention, the memory can be dynamically evolved along with the frequency and rule of user behaviors, the humanization degree of the system is continuously deepened along with time, the service meeting the current demand of the user is actively generated, memory grading and real-time context awareness can be combined, and the risk of privacy disclosure is effectively reduced.

The present application relates to a kind of data processing method and system for anonymous personnel attribute abundance optimization, it is related to data processing field, by obtaining first sequence to the privacy ordering of the investigation attribute set based on pre-defined privacy level table, obtain multiple groups of zero-knowledge feature value set using zero-knowledge processing data sandbox processing candidate personnel attribute characteristic value set, then according to the variance ordering of these feature value sets, obtain second sequence, combined with two sequences, sequence number is added and sorted to obtain third sequence, and finally according to third sequence, attribute abundance optimization is realized to realize personnel sorting, it solves the technical problem that it is difficult to balance screening efficiency and privacy protection when personnel screening is carried out to group in special scene, effectively balances screening efficiency and privacy protection, both improve the accuracy of personnel screening, and significantly reduce the risk of privacy leakage.

The invention relates to user device data privacy protection during machine learning model training. Techniques are disclosed for protecting privacy for user equipment during machine learning model training in a communication network environment. In one example, from the perspective of a user device, a method includes sending a selected privacy level from the user device to a communication network with which the communication network is participating in a machine learning model training process. The method also includes receiving, at the user device, one or more privacy parameters corresponding to the selected privacy level. The method further includes applying, by the user equipment, one or more privacy parameters to data to be sent to the communication network. In addition to the technical scheme of the user equipment side, the invention also provides a technical scheme of the communication network side, which is used for protecting the privacy of the data of the user equipment.