Privacy Data Clustering via Secret Sharing Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies lack a suitable method for clustering privacy data across multiple parties without leaking sensitive information, as they do not provide secure mechanisms for handling vertically distributed data.
Innovation Solution
A method and apparatus for clustering privacy data using iterative processes, where each party determines a part of the central data corresponding to a cluster, and joint comparison through secret sharing is used to determine the shortest target distance, ensuring that no single party can determine the central data alone, thereby preventing data leakage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If privacy data of multiple parties is clustered using existing technology, then clustering functionality is achieved, but privacy data leakage occurs
Solution Approach 1:
The patent segments the clustering process into multiple rounds of iterative calculations, where each party independently computes partial results (first segment, second segment) of the distance metric. These segments are then combined through secure multi-party computation to obtain the final distance without any single party accessing complete sensitive data. This segmentation prevents privacy leakage while maintaining clustering functionality.
Solution Approach 2:
The patent introduces secret sharing as an intermediary mechanism that enables joint comparison of distance segments without revealing the actual distance values or underlying private data. Each party holds shares of the distance segments, and through cryptographic protocols, they can determine the minimum distance without any party learning the complete distance information or other parties' private data.
2Productivity
If central data is determined by a single party, then calculation efficiency is improved, but data security is compromised
Solution Approach 1:
The patent divides the central data determination into collaborative segments where each party contributes their local data segments. The first party determines a first data part of central data corresponding to their feature dimensions, while the second party determines a second data part corresponding to their feature dimensions. This segmentation allows distributed computation efficiency while maintaining security through the fact that no single party possesses the complete central data.
3Ease of operation
If complete distance calculation is performed locally, then computational simplicity is achieved, but privacy protection is lost
Solution Approach 1:
The patent segments the distance calculation into multiple independent parts that can be computed locally and then combined securely. Each party calculates their segment of the distance metric using only their local data and the publicly available portions of central data, avoiding the need for complex secure computation throughout the entire distance calculation process.
Solution Approach 2:
The patent uses secret sharing as an intermediary to enable the combination of locally computed distance segments without revealing sensitive information. The joint comparison protocol acts as a mediator that allows parties to determine the minimum distance while protecting privacy, bridging the gap between local computational simplicity and centralized privacy protection.
Data Source
AI summary
Embodiments of the present specification provide a method and an apparatus for clustering privacy data of a plurality of parties. The method includes: a first party determining a first data part of central data currently corresponding to a cluster of K clusters, where the first data part corresponds to a first dimension set; a second party has a second data part of the central data corresponding to a second dimension set; separately using central data of a cluster of the K clusters as target central data, and obtaining a first segment of a target distance between a sample of the N samples and the target central data through local calculation based on a first feature part of any one of N samples and the first data part of the target central data; determining a shortest target distance among target distances by performing joint comparison using the first segment of each target distance and a second segment of each target distance of the second party through secret sharing; and determining a cluster corresponding to the shortest target distance as a cluster that any sample currently belongs to. As such, leakage of privacy data can be prevented.