Method and System for Increasing Privacy of User Data within a Dataset and Computer Program Product

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention provides a method, system, and computer program for increasing the privacy of user data in datasets containing information about multiple users, especially when users' data are correlated (e.g., family members, colleagues). It achieves this by transforming data to decorrelate it, applying advanced differential privacy mechanisms to the transformed data, and then reversing the transformation to yield a private dataset. This approach limits the possibility of inferring sensitive information about individuals, even when some relationships among users are known by potential attackers.

Use CasesContent extracted from patent full text and abstract with AI.

• Healthcare records: securely sharing or publishing anonymized patient datasets for research without exposing sensitive information.
• Retail and e-commerce analytics: protecting buyers' privacy when analyzing purchase patterns for marketing or business insights.
• Social network data sharing: enabling privacy-preserving analysis of user interactions that may be correlated (e.g., friends, family groups).
• Location and mobility data: sharing aggregated or anonymized movement data (e.g., from mobile phones or GPS) without individual tracing, even when users have overlapping paths.
• Internet of Things (IoT): protecting sensor data collected from multiple users/devices in smart homes or industrial networks.
• Federated learning: enhancing privacy guarantees in distributed machine learning where model parameters may leak user data.
• Government statistics: publishing census, crime, or employment data while maintaining confidentiality for individuals, especially in smaller or closely-knit communities.

BenefitsContent extracted from patent full text and abstract with AI.

• Stronger privacy guarantees even when user data are correlated, overcoming a key limitation of standard differential privacy approaches.
• Preservation of overall data utility—statistical patterns and aggregate information remain useful for analysis while individual information is protected.
• Adaptable to different datasets and privacy requirements by tuning parameters and choosing optimal transforms.
• Reduces the risk of inference attacks where attackers exploit known relationships between individuals to uncover private data.
• Supports centralized (server-side) privacy management without requiring end-users to handle privacy operations themselves, simplifying deployment in enterprise or cloud environments.
• Can be implemented with reasonable computational and hardware requirements, with mechanisms to discard overly complex methods.
• Applicable across a wide range of domains (health, finance, IoT, AI/ML) and data types (numerical, categorical, time-series, etc.).

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

A method for increasing privacy of user data of a plurality of users within a dataset is disclosed. The method comprises, in one or more data processing devices, providing (10) a dataset comprising a plurality of data points of a plurality of users and comprising inter-user correlations within the plurality of data points; determining (12) a plurality of transform coefficients by applying a transform on the plurality of data points; determining (14) a plurality of private transform coefficients from the plurality of transform coefficients by applying an (ε, δ)-differential privacy mechanism to each non-zero transform coefficient of the plurality of transform coefficients; and determining (15) a private dataset comprising a plurality of private data points from the plurality of private transform coefficients by applying, on the plurality of private transform coefficients, an inverse transform of the transform; wherein the (ε, δ)-differential privacy mechanism is adapted such that the plurality of private data points is (ε, δ)-differential private. Further, a system for increasing privacy of user data of a plurality of users within a dataset and a computer program product are provided.