Structure Preserving Encryption Network for Secure Analytics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional homomorphic encryption and differential privacy schemes are impractical due to increased resource requirements and larger key sizes, making them unsuitable for 'big data' analytics, and existing homomorphic encryption schemes are computationally inefficient and insecure.
Innovation Solution
Structure Preserving Encryption Networks (SPEN) transform plaintext data into ciphertext while preserving the data structure, using neural networks to enable efficient analytics without privacy leakage, with the addition of dummy dimensions for enhanced security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional homomorphic encryption schemes are used to protect data privacy, then data security is improved, but computational efficiency deteriorates and resource requirements increase
Solution Approach 1:
The patent changes the fundamental parameters of encryption by using structure-preserving encryption instead of traditional homomorphic encryption. This approach maintains multiplicative structure in the encrypted domain while using more practical key sizes and computational parameters, thereby improving computational efficiency without sacrificing security
Solution Approach 2:
The patent replaces the complex mechanical operations of traditional homomorphic encryption with a neural network-based encryption system. The SPEN uses learned transformations rather than rigid cryptographic operations, enabling more efficient computation while preserving data structure for analytics
2Reliability
If traditional homomorphic encryption schemes are used to protect data privacy, then data security is improved, but key size increases
Solution Approach 1:
The patent fundamentally changes the encryption parameter structure by using structure-preserving encryption with compact keys. The SPEN approach uses neural network weights as encryption keys, which can be much smaller than traditional homomorphic encryption keys while maintaining security through the complexity of the learned transformations
3Reliability
If data is encrypted for privacy protection, then data security is improved, but analytics capability deteriorates
Solution Approach 1:
The patent applies local quality by preserving specific structural properties (multiplicative structure) in the encrypted domain while transforming other aspects of the data. This allows analytics operations that rely on structure (like PCA and clustering) to function on encrypted data without requiring full decryption
Solution Approach 2:
The patent replaces traditional encryption mechanics with a neural network-based system that learns to preserve analytical structure. The SPEN transforms data in a way that maintains relationships needed for analytics while providing strong encryption, enabling both security and analytics capability simultaneously
Data Source
AI summary
A method for providing ciphertext data by a first computing device having memory includes obtaining, from the memory, plaintext data having a structure; providing the plaintext data to a structure preserving encryption network (SPEN) to generate the ciphertext data, where the structure of the plaintext data corresponds to a structure of the ciphertext data; and communicating, from the first computing device to a second computing device, the ciphertext data to permit analysis on the ciphertext data.