Secret Sharing With Trusted Initializer For Privacy-Preserving Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online platforms face challenges in collaborating to improve recommendation systems due to data privacy concerns, as existing methods require sharing of original user rating and social network data, compromising privacy and efficiency.
Innovation Solution
Implementing a secret sharing scheme with a trusted initializer allows platforms to securely collaborate by exchanging manipulated data, enabling improved recommendation models without disclosing private information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If platforms share original user rating and social network data to improve recommendation systems, then recommendation accuracy is improved, but data privacy is compromised
Solution Approach 1:
The patent segments the data collaboration process into multiple independent parties (social network platform, rating platform, and trusted initializer) who never directly share raw data. Each party holds and processes only their own data locally, breaking the traditional direct sharing model into distributed computational segments that preserve privacy while achieving collaborative recommendation improvement.
Solution Approach 2:
The patent introduces a trusted initializer as an intermediary that facilitates secure computation between platforms without having access to the actual private data. This intermediary enables the collaborative recommendation system to function by coordinating the secret sharing and computation processes while maintaining data confidentiality across all participating platforms.
2Loss of information
If platforms collaborate securely using secret sharing schemes, then data privacy is protected, but computational efficiency deteriorates
Solution Approach 1:
The patent performs preliminary data processing and feature extraction on each platform before the secure collaborative computation stage. By pre-computing user features, item features, and social network representations locally, the system reduces the complexity and computational burden of the subsequent secret sharing-based collaborative processing, thereby improving overall efficiency.
Solution Approach 2:
The patent extracts and separates the essential collaborative computation from the private data itself. By taking out only the necessary computational operations (matrix multiplications, dot products) and performing them through secret sharing on encrypted representations, the system avoids the overhead of processing entire datasets while maintaining privacy protection.
3Productivity
If platforms exchange manipulated data through trusted initializer, then collaboration efficiency is improved, but system complexity increases
Solution Approach 1:
The patent designs a universal secret sharing framework that can handle multiple types of data (social network data, rating data, user features, item features) and multiple computational operations (matrix multiplications, dot products, gradient computations) through a single unified protocol. This multi-functional approach simplifies the overall system architecture compared to implementing separate custom protocols for each data type and operation.
Solution Approach 2:
The system enables each platform to independently perform local computations and data transformations without requiring continuous coordination or complex communication protocols. Each platform self-manages its own secret sharing, local model training, and feature extraction, reducing the complexity of inter-platform coordination while maintaining collaborative efficiency.
Data Source
AI summary
An item rating and recommendation platform identifies rating data comprising respective ratings of multiple items with respect to multiple users, identifies user-feature data comprising multiple user features contributing to the respective ratings of the multiple items with respect to the multiple users, and receives, from a social network platform via a secret sharing scheme with a trusted initializer, manipulated social network data computed based on social network data and first input data from the trusted initializer. The social network data indicate social relationships between any two of the multiple users. In the secret sharing scheme with the trusted initializer, the social network platform shares with the item rating and recommendation platform the manipulated social network data without disclosing the social network data. The item rating and recommendation platform updates the user-feature data based on the rating data and the manipulated social network data.