Differential privacy protection method based on salient point privacy budget redistribution

Abstract

The invention relates to the technical field of differential privacy, and discloses a differential privacy protection method, device and equipment based on salient point privacy budget reallocation, and the method comprises the steps: obtaining the time series data of a user, and presetting the size of a sliding time window and a window-level privacy budget; dividing the time sequence data into a plurality of sliding time windows; uniformly distributing the window-level privacy budget to each time point in the sliding time window; identifying salient points and non-salient points in each sliding time window; the initial privacy budget corresponding to the non-salient point is recovered, the recovered initial privacy budget is allocated to the nearest salient point after the time point where the non-salient point is located, and the privacy budget of the salient point is updated; generating disturbance salient point data by using the updated privacy budget of the salient point; and reconstructing complete time sequence data by using a least square method according to the disturbance salient point data. The method can ensure that the reconstructed complete time sequence data meets the differential privacy protection requirement.

Description

Technical Field

[0001] This invention belongs to the field of differential privacy technology, and in particular relates to a differential privacy protection method based on salient point privacy budget redistribution. Background Technology

[0002] With the rapid development of big data technology, public and commercial decision-making is gradually shifting from static retrospective analysis of historical data to dynamic insights and real-time responses to continuous data streams. In the critical area of ​​personal health monitoring, real-time analysis of data streams and swift action have become core capabilities for improving social governance effectiveness and preventing systemic risks. However, while this decision-making paradigm based on real-time data streams unleashes enormous potential, its continuous and uninterrupted data collection and processing characteristics also expose individuals to unprecedented risks of continuous privacy exposure. Against the backdrop of increasingly stringent global privacy regulations (such as GDPR and CCPA) and growing public awareness of privacy, how to improve data usability while protecting user privacy has become a key challenge in unlocking the value of data.

[0003] Differential privacy (DP), with its rigorous mathematical definition and quantifiable protection guarantees, has become the theoretical cornerstone for protecting privacy in dynamic data environments. Its core advantage lies in its independence from attackers' background knowledge and its ability to provide an accurate accounting of cumulative privacy losses, which is crucial for streaming data processing scenarios that require the continuous publication of intermediate results. Traditional differential privacy mechanisms typically introduce carefully calibrated random noise to ensure that the exposure risk of a single data point, regardless of when it is published, is strictly controlled within a mathematically defined budget.

[0004] While differential privacy provides strong theoretical support, in real-world streaming data applications, the value and sensitivity of information often exhibit a significantly uneven distribution. Specifically, large amounts of continuously collected data often only reflect the normal operating state of the system, but show characteristics of small fluctuations and limited information increments; while data corresponding to a few key time points often reflect behavioral changes, abnormal states, or important events, possessing higher information density and potential privacy sensitivity. For example, in wearable device data streams, most of the data is in a stable state with low information entropy, with only a small number of segments containing abnormal events with high clinical diagnostic value (such as arrhythmia or a sudden drop in blood oxygen).

[0005] However, most existing differential privacy methods for streaming data employ static or uniform privacy budget allocation strategies, or budget allocation based on fixed decay models. This approach suffers from the following drawbacks: existing methods often ignore the temporal differences in data importance, treating the privacy budget as a "one-time consumable" and distributing it uniformly within a fixed time window. This coarse-grained allocation leads to a significant waste of valuable privacy budget on ordinary data points with limited information value. Furthermore, because the budget is evenly distributed, when truly critical data points (such as anomalous medical events) occur, insufficient available budget in certain areas often forces the system to add excessive noise to meet privacy constraints. This masks salient information, potentially resulting in lower usability of time-series data and lower accuracy in subsequent data analysis. Summary of the Invention

[0006] This invention provides a differential privacy protection method based on salient point privacy budget reallocation, which can ensure that the reconstructed complete time series data meets the differential privacy protection requirements.

[0007] To achieve the above objectives, this invention provides a differential privacy protection method based on saliency point privacy budget reallocation, comprising: Acquire users' time-series data and preset the sliding time window size and the corresponding window-level privacy budget. The time-series data is personal health monitoring data. Divide the time series data into multiple sliding time windows based on the size of the sliding time window; The window-level privacy budget is evenly distributed to each time point within the sliding time window to obtain the initial privacy budget for each time point; Within each sliding time window, significant and insignificant points are identified based on the data change trends at adjacent time points; The initial privacy budget corresponding to the non-salient point is reclaimed, and the reclaimed initial privacy budget is allocated to the most recent salient point after the time point of the non-salient point. The privacy budget of the salient point is then updated using the reclaimed initial privacy budget. Using the updated privacy budget for salient points, local differential privacy perturbation is performed on the salient point data within the sliding time window to generate perturbed salient point data; Based on the data of significant disturbance points, the complete time series data is reconstructed using the least squares method.

[0008] To address the aforementioned problems, the present invention also provides a differential privacy protection device based on salient point privacy budget reallocation, the device comprising: The time series data segmentation module is used to acquire users' time series data and preset the sliding time window size and the corresponding window-level privacy budget. The time series data is personal health monitoring data. The time series data is divided into multiple sliding time windows according to the sliding time window size. The privacy budget allocation module is used to evenly distribute the window-level privacy budget to each time point within the sliding time window to obtain the initial privacy budget for each time point; within each sliding time window, it identifies salient and insignificant points based on the data change trends of adjacent time points; it reclaims the initial privacy budget corresponding to the insignificant points and allocates the reclaimed initial privacy budget to the nearest salient point after the time point where the insignificant point is located, and updates the privacy budget of the salient points using the reclaimed initial privacy budget; The time series data reconstruction module is used to perform local differential privacy perturbation on the salient point data within the sliding time window using the updated privacy budget of the salient points, generating perturbed salient point data; and reconstructing the complete time series data using the least squares method based on the perturbed salient point data.

[0009] To address the above problems, the present invention also provides an electronic device, the electronic device comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the differential privacy protection method based on salient point privacy budget reallocation described above.

[0010] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one computer program, which is executed by a processor in an electronic device to implement the differential privacy protection method based on salient point privacy budget reallocation described above.

[0011] This invention acquires users' time-series data, presets a sliding time window size and a corresponding window-level privacy budget, and divides the time-series data into multiple sliding time windows based on the sliding time window size. By introducing a sliding time window mechanism, infinite time-series data can be transformed into finite and measurable processing units, ensuring the controllability of privacy protection in the time dimension. Furthermore, evenly distributing the window-level privacy budget to each time point within the sliding time window yields an initial privacy budget for each time point, ensuring that each time point has a privacy budget and increasing the importance of subsequent redistribution of the initial privacy budget to salient points. Moreover, within each sliding time window, identifying salient and non-salient points based on the data change trends of adjacent time points ensures the rationality of the privacy budget allocation at the value level. Additionally, the initial privacy budget corresponding to non-salient points is... Budget recovery and allocation of the recovered initial privacy budget to the most recent salient point after the time point of the non-salient point, followed by updating the privacy budget of the salient points using the recovered initial privacy budget, enables stronger privacy protection for salient points. Finally, using the updated privacy budget of the salient points, local differential privacy perturbation is performed on the salient point data within the sliding time window to generate perturbed salient point data. Based on the perturbed salient point data, the complete time series data is reconstructed using the least squares method. By performing least squares fitting on the perturbed salient point data, the complete time series data is reconstructed, so that non-salient points do not need to directly publish the original data or be perturbed separately, but are inferred through fitting the salient points. This ensures that the reconstructed complete time series data meets the differential privacy protection requirements and further improves the accuracy of subsequent data analysis using the reconstructed complete time series data. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating a differential privacy protection method based on salient point privacy budget reallocation, provided in an embodiment of the present invention. Figure 2 This is a functional block diagram of a differential privacy protection device based on salient point privacy budget redistribution according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device that implements the differential privacy protection method based on salient point privacy budget redistribution, as provided in an embodiment of the present invention.

[0013] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0014] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0015] This application provides a differential privacy protection method based on saliency-based privacy budget reallocation. The executing entity of this differential privacy protection method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the differential privacy protection method based on saliency-based privacy budget reallocation can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0016] Reference Figure 1 The diagram shown is a flowchart illustrating a differential privacy protection method based on saliency point privacy budget reallocation according to an embodiment of the present invention. In this embodiment, the differential privacy protection method based on saliency point privacy budget reallocation includes: S1. Obtain the user's time series data and preset the sliding time window size and the window-level privacy budget corresponding to the sliding time window size. The time series data is personal health monitoring data.

[0017] Understandably, a user's time series data refers to a set of data points collected continuously in chronological order. For example, a user's time series data could be a data stream such as heart rate and blood oxygen in personal health monitoring data.

[0018] Understandably, a sliding time window refers to a method of dividing continuous time series data into segments of fixed length. The data processing mechanism for fragmented data.

[0019] For example, given time series data Define a sliding window of length w as follows: Each sliding window is considered an independent time event interval, called a w-event. ,in For each time point, each sliding time window It is treated as an independent privacy protection unit for time event intervals, and subsequent salient point identification, privacy budget allocation and perturbation operations are all carried out independently within the sliding time window.

[0020] Understandably, window-level privacy budget refers to the differential privacy budget ε allocated to each sliding time window, which is used to control the strength of privacy protection for data within that window.

[0021] For example, window-level privacy budgeting characterizes the privacy budget in any continuous Within each specific timeframe, information leaked due to data changes is limited to a window-level privacy budget. Within, among which parameters It is the size of a sliding time window, representing a local time scale. The smaller the value, the more pronounced transient anomalies become; it requires that the value be within any continuous range. Within each time point, the total accumulated privacy breaches do not exceed the window-level privacy budget. This allows for more flexible allocation of privacy budgets over time.

[0022] S2. Divide the time series data into multiple sliding time windows according to the size of the sliding time window.

[0023] S3. Distribute the window-level privacy budget evenly to each time point within the sliding time window to obtain the initial privacy budget for each time point.

[0024] Specifically, the window-level privacy budget is evenly distributed to each time point within the sliding time window to obtain the initial privacy budget for each time point, including: A fixed window-level privacy budget is allocated to the sliding time window, and the window-level privacy budget is evenly distributed to each time point within the sliding time window using a preset privacy budget allocation formula to obtain the initial privacy budget for each time point. The preset privacy budget allocation formula is as follows: ,in, For the first The initial privacy budget for each sliding time window. For window-level privacy budget, This represents the number of time points within the sliding time window.

[0025] S4. Within each sliding time window, identify significant and insignificant points based on the data change trends of adjacent time points.

[0026] Understandably, salient points refer to key data points that can reflect dramatic changes or trend reversals (such as peaks, troughs, and inflection points) within a local area.

[0027] Understandably, non-significant points refer to data points that change steadily and contain little information.

[0028] Specifically, significant and insignificant points are identified based on the data change trends at adjacent time points, including: Calculate the first difference between data points at adjacent time points: ,in, For time points Data values, For time points The data value at the previous time point, For time points The first-order difference; If the conditions are met Then determine the time point. As salient points, among which, For time points The first difference at the previous time point; If the conditions are not met Then determine the time point. This is a non-significant point.

[0029] It is understandable that the identification of salient and non-salient points can also be constructed based on one or more of the following features, including but not limited to: the magnitude of change in the time series, the direction of change, the rate of change, local extremum features, differences or higher-order differences, and statistical change thresholds.

[0030] S5. Reclaim the initial privacy budget corresponding to the non-salient point, allocate the reclaimed initial privacy budget to the most recent salient point after the time point of the non-salient point, and update the privacy budget of the salient point using the reclaimed initial privacy budget.

[0031] Specifically, the privacy budget for salient points is updated using the recovered initial privacy budget, including: The privacy budget for salient points is updated using the following formula: ; in, For the first Privacy budget for salient points updated within a sliding time window. For the first The first sliding time window The initial privacy budget that is recovered at each time step For the first The initial privacy budget for salient points that have not been updated within a sliding time window.

[0032] Understandably, embodiments of the present invention reclaim the initial privacy budget corresponding to non-salient points and allocate the reclaimed initial privacy budget to the most recent salient point after the time point where the non-salient point is located, which can prevent the privacy budget from being consumed on low-value data.

[0033] S6. Using the updated privacy budget for salient points, perform local differential privacy perturbation on the salient point data within the sliding time window to generate perturbed salient point data.

[0034] Understandably, local differential privacy refers to a decentralized privacy protection technology, such as adding random noise that conforms to a Laplace distribution to user data before it leaves the local device.

[0035] For example, a differential privacy mechanism is defined as follows: suppose a random mechanism M is applied to the entire time series, and any event-level adjacent data sequences as defined above... and And any output set O, all satisfy: The above mechanism is said to satisfy time-event interval differential privacy, where, For time series data, For event-level adjacent time series data, Given time series data X as input, the random mechanism M outputs the probability that the set belongs to set O. When the input is event-level adjacent time series data When the random mechanism M outputs the probability that set M belongs to set O, This serves as an upper bound for privacy breaches. Furthermore, if the randomization process for local differential privacy occurs locally on the user's device, meaning the time-series data is randomized before uploading, then the mechanism is said to satisfy local differential privacy for time event intervals.

[0036] Specifically, local differential privacy perturbations are performed on salient point data within the sliding time window, including: The following Laplace mechanism is used to perform local difference privacy perturbations on salient point data: ; in, This is data on significant perturbation points after local differential privacy perturbation. For salient point data, For global sensitivity, For the first Privacy budget for significant points after a sliding time window update.

[0037] Understandably, global sensitivity is defined as follows: Let the function... A function that performs calculations on input data, such as mean or median calculations, outputs a real value or a vector. Global sensitivity measures the maximum possible change in a function's output when data at a single point in time changes; it is defined as: ;in and Any two adjacent input data, It measures the maximum change that a query result can undergo when any single user adds or removes their data from the data stream at any possible point in time. Global sensitivity reflects how sensitive the function is to changes in the input and is an important parameter for calibrating the intensity of differential privacy noise.

[0038] Understandably, adjacent inputs are defined as follows: In streaming data / time series scenarios, two sets of input data... and They are called event-level adjacencies if and only if, at a single point in time, the data about the same user differs (e.g., at that point...). Includes user data (Not included, or the data values ​​are different), but the data is exactly the same at all other points in time.

[0039] S7. Based on the data of significant disturbance points, reconstruct the complete time series data using the least squares method.

[0040] Understandably, the least squares method is a mathematical optimization algorithm that can use the received sparse, noisy salient point data to construct an objective function (minimizing the sum of squared errors) to reverse-engineer the original complete curve, and can solve the data missing problem caused by only uploading salient points.

[0041] Specifically, based on the data of significant disturbance points, the complete time series data is reconstructed using the least squares method, including: The complete time series data is reconstructed using the following least squares objective function: ; in, Here is the regularized least squares estimation function. For the reconstructed complete time series data, For smoothing weight parameters, For significant point time indexes, These are estimates of the significance points. For the perturbation observations at significant points, This represents the total number of time points in the time series data. For time points The estimated value, For time points The estimated value.

[0042] This invention acquires users' time-series data, presets a sliding time window size and a corresponding window-level privacy budget, and divides the time-series data into multiple sliding time windows based on the sliding time window size. By introducing a sliding time window mechanism, infinite time-series data can be transformed into finite and measurable processing units, ensuring the controllability of privacy protection in the time dimension. Furthermore, evenly distributing the window-level privacy budget to each time point within the sliding time window yields an initial privacy budget for each time point, ensuring that each time point has a privacy budget and increasing the importance of subsequent redistribution of the initial privacy budget to salient points. Moreover, within each sliding time window, identifying salient and non-salient points based on the data change trends of adjacent time points ensures the rationality of the privacy budget allocation at the value level. Additionally, the initial privacy budget corresponding to non-salient points is... Budget recovery and allocation of the recovered initial privacy budget to the most recent salient point after the time point of the non-salient point, followed by updating the privacy budget of the salient points using the recovered initial privacy budget, enables stronger privacy protection for salient points. Finally, using the updated privacy budget of the salient points, local differential privacy perturbation is performed on the salient point data within the sliding time window to generate perturbed salient point data. Based on the perturbed salient point data, the complete time series data is reconstructed using the least squares method. By performing least squares fitting on the perturbed salient point data, the complete time series data is reconstructed, so that non-salient points do not need to directly publish the original data or be perturbed separately, but are inferred through fitting the salient points. This ensures that the reconstructed complete time series data meets the differential privacy protection requirements and further improves the accuracy of subsequent data analysis using the reconstructed complete time series data.

[0043] like Figure 2 The diagram shown is a functional block diagram of a differential privacy protection device based on salient point privacy budget redistribution provided in an embodiment of the present invention.

[0044] The differential privacy protection device 100 based on salient point privacy budget redistribution described in this invention can be installed in an electronic device. Depending on the functions implemented, the differential privacy protection device 100 may include a time series data partitioning module 101, a privacy budget allocation module 102, and a time series data reconstruction module 103.

[0045] The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0046] In this embodiment, the functions of each module / unit are as follows: The time series data partitioning module 101 is used to acquire the user's time series data and preset the sliding time window size and the window-level privacy budget corresponding to the sliding time window size, wherein the time series data is personal health monitoring data; the time series data is divided into multiple sliding time windows according to the sliding time window size.

[0047] The privacy budget allocation module 102 is used to evenly distribute the window-level privacy budget to each time point within the sliding time window to obtain the initial privacy budget for each time point; within each sliding time window, it identifies salient and insignificant points based on the data change trends of adjacent time points; it reclaims the initial privacy budget corresponding to the insignificant point, and allocates the reclaimed initial privacy budget to the nearest salient point after the time point where the insignificant point is located, and updates the privacy budget of the salient point using the reclaimed initial privacy budget.

[0048] The time series data reconstruction module 103 is used to perform local differential privacy perturbation on the salient point data within the sliding time window using the updated privacy budget of the salient points to generate perturbed salient point data; and to reconstruct the complete time series data using the least squares method based on the perturbed salient point data.

[0049] like Figure 3 The diagram shown is a structural schematic of an electronic device that implements a differential privacy protection method based on salient point privacy budget redistribution, according to an embodiment of the present invention.

[0050] The electronic device may include a processor 10, a memory 11, a communication bus 12, and a communication interface 13. It may also include a computer program stored in the memory 11 and executable on the processor 10, such as a differential privacy protection method program based on salient point privacy budget reallocation.

[0051] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., executing a differential privacy protection method program based on salient point privacy budget reallocation), and calls data stored in the memory 11 to perform various functions of the electronic device and process data.

[0052] The memory 11 includes at least one type of readable storage medium, including flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of an electronic device, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device, such as a plug-in portable hard drive, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc. Furthermore, the memory 11 can include both internal and external storage units of the electronic device. The memory 11 can be used not only to store application software and various types of data installed on the electronic device, such as the code of a differential privacy protection method program based on salient point privacy budget redistribution, but also to temporarily store data that has been output or will be output.

[0053] The communication bus 12 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory 11 and at least one processor 10, etc.

[0054] The communication interface 13 is used for communication between the aforementioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, Bluetooth interface, etc.), typically used to establish communication connections between the electronic device and other electronic devices. The user interface may be a display, an input unit (such as a keyboard), or optionally, a standard wired or wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen, etc. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device and to display a visual user interface.

[0055] Figure 3 Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 3The structure shown does not constitute a limitation on the electronic device and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0056] For example, although not shown, the electronic device may also include a power supply (such as a battery) to power the various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0057] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0058] The program for a differential privacy protection method based on salient point privacy budget reallocation, stored in the memory 11 of the electronic device, is a combination of multiple instructions that, when run in the processor 10, can achieve the following: Acquire users' time-series data and preset the sliding time window size and the corresponding window-level privacy budget. The time-series data is personal health monitoring data. Divide the time series data into multiple sliding time windows based on the size of the sliding time window; The window-level privacy budget is evenly distributed to each time point within the sliding time window to obtain the initial privacy budget for each time point; Within each sliding time window, significant and insignificant points are identified based on the data change trends at adjacent time points; The initial privacy budget corresponding to the non-salient point is reclaimed, and the reclaimed initial privacy budget is allocated to the most recent salient point after the time point of the non-salient point. The privacy budget of the salient point is then updated using the reclaimed initial privacy budget. Using the updated privacy budget for salient points, local differential privacy perturbation is performed on the salient point data within the sliding time window to generate perturbed salient point data; Based on the data of significant disturbance points, the complete time series data is reconstructed using the least squares method.

[0059] Specifically, the specific implementation method of the processor 10 for the above instructions can be referred to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, and will not be repeated here.

[0060] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0061] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following: Acquire users' time-series data and preset the sliding time window size and the corresponding window-level privacy budget. The time-series data is personal health monitoring data. Divide the time series data into multiple sliding time windows based on the size of the sliding time window; The window-level privacy budget is evenly distributed to each time point within the sliding time window to obtain the initial privacy budget for each time point; Within each sliding time window, significant and insignificant points are identified based on the data change trends at adjacent time points; The initial privacy budget corresponding to the non-salient point is reclaimed, and the reclaimed initial privacy budget is allocated to the most recent salient point after the time point of the non-salient point. The privacy budget of the salient point is then updated using the reclaimed initial privacy budget. Using the updated privacy budget for salient points, local differential privacy perturbation is performed on the salient point data within the sliding time window to generate perturbed salient point data; Based on the data of significant disturbance points, the complete time series data is reconstructed using the least squares method.

[0062] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0063] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0064] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0065] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0066] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0067] The blockchain referred to in this invention is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0068] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0069] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in the system claims may also be implemented by a single unit or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any specific order.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A differential privacy protection method based on salient point privacy budget reallocation, characterized in that, The method includes: Acquire users' time-series data and preset the sliding time window size and the corresponding window-level privacy budget. The time-series data is personal health monitoring data. Divide the time series data into multiple sliding time windows based on the size of the sliding time window; The window-level privacy budget is evenly distributed to each time point within the sliding time window to obtain the initial privacy budget for each time point; Within each sliding time window, significant and insignificant points are identified based on the data change trends at adjacent time points; The initial privacy budget corresponding to the non-salient point is reclaimed, and the reclaimed initial privacy budget is allocated to the most recent salient point after the time point of the non-salient point. The privacy budget of the salient point is then updated using the reclaimed initial privacy budget. Using the updated privacy budget for salient points, local differential privacy perturbation is performed on the salient point data within the sliding time window to generate perturbed salient point data; Based on the data of significant disturbance points, the complete time series data is reconstructed using the least squares method.

2. The differential privacy protection method based on salient point privacy budget reallocation as described in claim 1, characterized in that, The method of identifying significant and insignificant points based on the data change trends of adjacent time points includes: Calculate the first difference between data points at adjacent time points: ,in, For time points Data values, For time points The data value at the previous time point, For time points The first-order difference; If the conditions are met Then determine the time point. As salient points, among which, For time points The first difference at the previous time point; If the conditions are not met Then determine the time point. This is a non-significant point.

3. The differential privacy protection method based on salient point privacy budget reallocation as described in claim 1, characterized in that, The process of performing local differential privacy perturbation on salient point data within the sliding time window includes: The following Laplace mechanism is used to perform local difference privacy perturbations on salient point data: ； in, This is data on significant perturbation points after local differential privacy perturbation. For salient point data, For global sensitivity, For the first Privacy budget for significant points after a sliding time window update.

4. The differential privacy protection method based on salient point privacy budget reallocation as described in claim 1, characterized in that, The process of reconstructing complete time series data using the least squares method based on significant perturbation data includes: The complete time series data is reconstructed using the following least squares objective function: ； in, Here is the regularized least squares estimation function. For the reconstructed complete time series data, For smoothing weight parameters, For significant point time indexes, These are estimates of the significance points. For the perturbation observations at significant points, This represents the total number of time points in the time series data. For time points The estimated value, For time points The estimated value.

5. The differential privacy protection method based on salient point privacy budget reallocation as described in claim 1, characterized in that, The process of evenly distributing the window-level privacy budget to each time point within the sliding time window to obtain the initial privacy budget for each time point includes: A fixed window-level privacy budget is allocated to the sliding time window, and the window-level privacy budget is evenly distributed to each time point within the sliding time window using a preset privacy budget allocation formula to obtain the initial privacy budget for each time point. The preset privacy budget allocation formula is as follows: ,in, For the first The initial privacy budget for each sliding time window. For window-level privacy budget, This represents the number of time points within the sliding time window.

6. The differential privacy protection method based on salient point privacy budget reallocation as described in claim 1, characterized in that, The process of updating the privacy budget for salient points using the recovered initial privacy budget includes: The privacy budget for salient points is updated using the following formula: ； in, For the first Privacy budget for salient points updated within a sliding time window. For the first The first sliding time window The initial privacy budget that is recovered at each time step For the first The initial privacy budget for salient points that have not been updated within a sliding time window.

7. A differential privacy protection device based on salient point privacy budget reallocation, characterized in that, The apparatus is used to implement the differential privacy protection method based on salient point privacy budget reallocation as described in any one of claims 1 to 6, the apparatus comprising: The time series data segmentation module is used to acquire users' time series data and preset the sliding time window size and the corresponding window-level privacy budget. The time series data is personal health monitoring data. The time series data is divided into multiple sliding time windows according to the sliding time window size. The privacy budget allocation module is used to evenly distribute the window-level privacy budget to each time point within the sliding time window to obtain the initial privacy budget for each time point; within each sliding time window, it identifies salient and insignificant points based on the data change trends of adjacent time points; it reclaims the initial privacy budget corresponding to the insignificant points and allocates the reclaimed initial privacy budget to the nearest salient point after the time point where the insignificant point is located, and updates the privacy budget of the salient points using the reclaimed initial privacy budget; The time series data reconstruction module is used to perform local differential privacy perturbation on the salient point data within the sliding time window using the updated privacy budget of the salient points, generating perturbed salient point data; and reconstructing the complete time series data using the least squares method based on the perturbed salient point data.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor to enable the at least one processor to perform the differential privacy protection method based on salient point privacy budget reallocation as described in any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the differential privacy protection method based on salient point privacy budget reallocation as described in any one of claims 1 to 6.

Images