Homomorphic Encryption Automorphism Blind Rotation Key Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing homomorphic encryption schemes face challenges with large public key sizes and increased computational requirements due to the size of the public key, which hampers efficiency and security in operations.
Innovation Solution
The proposed solution involves generating a blind rotation key based on automorphism, preprocessing the operand ciphertext, and performing blind rotation operations using increment, automorphism, and key switching operations to reduce the size of the public key and computational overhead, while maintaining high security through probability distributions like Gaussian distributions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If blind rotation operation technology is used to apply arbitrary function operations to ciphertext messages, then operation accuracy is improved, but public key size increases significantly
Solution Approach 1:
The patent segments the blind rotation operation into multiple components: preprocessing the ciphertext to extract vector components, separating the public key into basis vectors, and performing rotations through component-wise operations. This segmentation allows accurate blind rotation without requiring a single large public key
Solution Approach 2:
The patent transforms the problem from operating on entire ciphertext vectors to operating on individual vector components in a higher-dimensional space. By preprocessing the ciphertext to obtain components and performing rotations component-wise, the system achieves accurate arbitrary function operations without proportionally increasing public key size
2Adaptability or versatility
If the size of public key is increased to enable arbitrary function operations, then operational capability is improved, but computational requirements increase significantly
Solution Approach 1:
The patent performs preliminary preprocessing on the ciphertext to extract vector components before the actual blind rotation operation. This preliminary action organizes the data structure so that subsequent rotations can be performed efficiently on components rather than on the entire large ciphertext, reducing computational requirements
Solution Approach 2:
Instead of performing full blind rotation on the entire ciphertext at once, the patent applies partial rotations to individual vector components. This partial action approach enables arbitrary function operations through multiple smaller, more efficient component-wise operations rather than one large computationally intensive operation
3Adaptability or versatility
If various blind rotation operation technologies are used, then functional versatility is improved, but memory requirements increase significantly
Solution Approach 1:
The patent segments the ciphertext into vector components through preprocessing, allowing each component to be handled separately in memory. This segmentation reduces the memory footprint compared to storing and manipulating entire large ciphertext structures for various blind rotation operations
Solution Approach 2:
The patent works with copied vector components extracted from the original ciphertext rather than manipulating the full ciphertext structure. These component copies are smaller in size and can be processed efficiently in memory, enabling functional versatility without proportionally increasing memory requirements
Data Source
AI summary
Disclosed are an apparatus and method with homomorphic encryption using automorphism. A computing apparatus includes one or more processors and a memory storing instructions configured to cause the one or more processors to, for a blind rotation key for performing a blind rotation operation and an operand ciphertext of the blind rotation operation: generate a preprocessed ciphertext by performing preprocessing on the operand ciphertext based on automorphism, and generate an operation result of the homomorphic encryption by performing the blind rotation operation for the operand ciphertext on a vector component of the preprocessed ciphertext and a vector component of the blind rotation key.