Connection Recommendations That Obscure Social Graph Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating meaningful connection recommendations for new users in a messaging system at registration time while preserving user privacy is challenging due to limited available information and potential leakage of user details.
Innovation Solution
A connection recommendation system that generates recommendations by selecting profiles connected to contact book matches, obscuring connection sources and distances, and maintaining a balanced ratio between different categories of profiles, using a hybrid graph approach with one-hop, two-hop, and three-hop profiles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If connection recommendations are generated using contact book matches and social graph data, then the quality and meaningfulness of recommendations improves, but user privacy is compromised due to potential leakage of connection details and distances
Solution Approach 1:
The patent extracts only the necessary connection information (that users share common connections) while leaving out sensitive details (connection distances, specific path information). The system recommends users based on shared connections without revealing how many hops separate them or the structure of the connection path, thus maintaining recommendation quality while protecting privacy.
Solution Approach 2:
The patent introduces an intermediary processing layer that receives raw social graph data, processes it to remove sensitive information, and outputs sanitized recommendation data. This intermediary layer transforms the connection data to preserve utility for recommendations while eliminating privacy-sensitive elements like connection distances and detailed path information.
2Ease of operation
If connection recommendations are generated at registration time with limited user information, then user experience is improved by providing immediate recommendations, but recommendation accuracy deteriorates due to insufficient data
Solution Approach 1:
The patent performs preliminary connection recommendation generation at registration time using available contact book information and initial social graph data. By preparing and providing recommendations immediately during registration, the system improves user experience without waiting for additional user interactions to accumulate more data.
3Measurement precision
If detailed connection information is provided in recommendations, then recommendation precision is improved, but information security is compromised
Solution Approach 1:
The patent extracts and retains only the essential fact that users share common connections, while removing sensitive details such as connection distances (one-hop, two-hop, three-hop relationships) and specific path information. This selective extraction maintains sufficient precision for meaningful recommendations while eliminating security-risk information.
Data Source
AI summary
When a messaging system generates connection recommendations for a new user, who first registers with the messaging system, the signals available for generation of recommendations may be limited to the user's contact book matches. Using just this limited signal poses a concern associated with leaking information about users represented by the recommendations. The technical problem of generating connection recommendations for a user at registration time in a privacy-safe manner is addressed by a recommendation methodology that obscures the connection source and the connection distance of the recommended profiles.