MPC User Labeling Without Third-Party Cookies or Data Exposure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data processing technologies face challenges in grouping user data across multiple websites without using third-party cookies, while ensuring user privacy and security, especially in environments where browsers block such cookies.
Innovation Solution
Implementing secure multi-party computation (MPC) techniques to split and encrypt user profiles across multiple computing systems, allowing them to determine labels for user clusters without revealing the complete dataset, and using exponential decay to reduce data size and storage requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user profiles are shared across multiple computing systems for machine learning model training, then model accuracy and user clustering capability are improved, but user privacy and data security are compromised
Solution Approach 1:
The user profile is divided into multiple secret shares distributed across different computing systems. Each system receives only a portion of the data (first share, second share, etc.), making it impossible to reconstruct the complete user profile without collusion. This segmentation enables collaborative machine learning model training while preserving user privacy, as no single system has access to the full dataset.
2Measurement precision
If complete user profiles are transmitted to computing systems for processing, then processing accuracy is improved, but data transmission security and storage security are worsened
Solution Approach 1:
User profiles are segmented into multiple secret shares that are transmitted separately to different computing systems. The segmentation ensures that even if transmission channels are compromised, the complete user profile cannot be reconstructed from partial shares alone, thereby maintaining data transmission security while enabling accurate processing through collaborative computation.
Solution Approach 2:
A multi-party computation protocol acts as an intermediary mechanism that enables secure collaboration between computing systems. The protocol allows systems to jointly process encrypted data and generate accurate results without any single system having access to the complete unencrypted user profile, thus maintaining both processing accuracy and transmission security.
3Adaptability or versatility
If third-party cookies are used for user grouping, then user data aggregation capability is improved, but system complexity and dependency on external systems increase
Solution Approach 1:
The system performs user profiling and clustering operations autonomously using secret-shared data from multiple sources, eliminating dependency on third-party cookies. The multi-party computation system self-manages the aggregation of user data across computing systems, providing user grouping capability without external cookie-based infrastructure, thereby reducing system complexity and external dependencies.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for identifying labels for a dataset without revealing the dataset to any individual computing system. Methods can include receiving, by a first computing system of a multi-party computation (MPC) system, a query that includes a first and second share of a given user profile. The second share is encrypted with a key that prevents the first computing system from accessing the second share. The second share is transmitted to a second computing system of the MPC system. The first and the second computing system generates a machine learning model and identifies a respective first and a second label. The first computing system receives the second label as a response from the second computing system. The first computing system responds to the query with a response that includes the first and the second label.