Third-Party User Group Identification via Obfuscated Bloom Filter Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Third-party systems face limitations in obtaining user information from online systems without revealing the group composition, which can lead to competition with the online system's content and privacy concerns.
Innovation Solution
A third-party system generates a function, such as a bloom filter, to identify users within a group without disclosing the group membership, allowing it to obtain information from the online system while preventing the online system from reproducing the group, and obfuscates user identifying information to maintain privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the third party system provides user identifying information associated with each user of the group to the online system, then the third party system can leverage information associated with the group of users by the online system, but the online system can reproduce the group and provide competing content to the users
Solution Approach 1:
The patent extracts only the necessary information (user properties and group statistics) from the group composition, allowing the third party system to obtain useful user information without revealing the actual group membership to the online system. This prevents the online system from reproducing the group while still enabling content customization.
Solution Approach 2:
The patent introduces an intermediary mechanism where the third party system processes user identifying information locally to generate group identifiers and statistics, which are then shared with the online system without exposing individual user memberships. This intermediary layer prevents direct group reproduction by the online system.
2Measurement precision
If the third party system obtains detailed user information from the online system, then content selection and user analysis can be improved, but user privacy is compromised
Solution Approach 1:
The patent applies local quality by allowing the third party system to access detailed user information only for the specific purpose of generating group statistics and selecting content, while the online system maintains control over the actual user data. User privacy is preserved because the online system does not receive or store individual user identifying information.
Solution Approach 2:
The patent extracts only the necessary aggregated information (group identifiers and statistics) from the user data, enabling precise user analysis and content selection without compromising individual user privacy. The third party system performs local processing to derive insights without exposing raw user information.
Data Source
AI summary
A third party system generates a group of users and a function that identifies users in the group as well as additional users not in the group when applied to user identifying information. The third party system transmits the function to an online system, which applies the function to user identifying information associated with various users of the online system. Applying the function to the user identifying information generates a set of users including users in the group and one or more additional users who are not in the group. The online system transmits information associated with users in the set and information identifying users in the set to the third party system, which determines obtained information associated with users of the group. In some embodiments, the information identifying users in the set is obfuscated user identifying information associated with the users in the set by the online system.