Bloom Filter Union Estimation for Privacy-Safe Audience Cardinality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audience measurement systems face challenges in accurately estimating the cardinality of users across multiple bloom filter arrays due to privacy concerns and the inability to share personally identifiable information, leading to double counting and inaccurate audience size estimates.
Innovation Solution
The use of Bloom filter arrays with agreed-upon hash functions to generate sketch data that preserves user privacy by ensuring that the same individuals are not double-counted across different datasets, allowing for accurate estimation of unique audience size without revealing personal identifiable information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If database proprietors share user data across multiple platforms to improve audience measurement accuracy, then measurement precision improves, but user privacy is compromised
Solution Approach 1:
The patent extracts only the necessary identifying information (user identifiers and exposure data) from complete user profiles, allowing audience measurement without sharing sensitive personal data. Database proprietors send minimal sketch data to the AME, eliminating the need to share full user records while maintaining measurement capability.
Solution Approach 2:
The patent introduces an intermediary audience measurement entity (AME) that receives anonymized sketch data from multiple database proprietors and performs centralized deduplication. This mediator enables cross-platform audience measurement without requiring direct data sharing between proprietors, preserving privacy while improving accuracy.
2Object-affected harmful factors
If traditional panel-based audience measurement is used to protect user privacy, then privacy is preserved, but measurement precision deteriorates due to limited panel coverage
Solution Approach 1:
The patent creates anonymized copies (sketch data) of user exposure information that preserve measurement capability without revealing identity. These copies contain user identifiers and exposure data sufficient for deduplication and audience calculation, but lack the granularity to identify individuals, enabling both privacy protection and improved measurement through broader data collection.
3Measurement precision
If complete user data is collected across multiple database proprietors to eliminate double counting, then measurement precision improves, but device complexity and processing demands increase
Solution Approach 1:
The patent extracts only essential identifying fields (user ID and exposure indicators) from complete user datasets, creating compact sketch data structures. This extraction reduces data volume and processing requirements while retaining sufficient information for accurate deduplication and unique audience calculation across multiple proprietors.
4Measurement precision
If third-party cookies are used to track users across platforms to improve measurement accuracy, then measurement precision improves, but ease of operation deteriorates due to browser restrictions and privacy concerns
Solution Approach 1:
The patent replaces third-party cookies with an intermediary AME that receives anonymized sketch data from database proprietors and performs centralized matching. This approach achieves cross-platform tracking without relying on browser cookie mechanisms, avoiding restrictions on third-party cookies while maintaining measurement capability through server-side processing.
Data Source
AI summary
Methods and apparatus to estimate cardinality of users represented across multiple bloom filter arrays are disclosed. Examples includes processor circuitry to execute and/or instantiate instructions to generate a first composite Bloom filter array based on first and second Bloom filter arrays. The processor circuitry is to generate a final composite Bloom filter array based on the first composite Bloom filter array and a third Bloom filter array. Different ones of the first, second, and third Bloom filter arrays representative of different sets of users who accessed media. The first, second, and third Bloom filter arrays including differential privacy noise. The processor circuitry to estimate a cardinality of a union of the first, second, and third Bloom filter arrays based on the final composite Bloom filter array.