Homomorphic Encryption Bootstrapping Using Chebyshev Polynomials

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current fully homomorphic encryption methods face challenges in efficiently bootstrapping ciphertexts due to noise accumulation, which requires effective noise processing to maintain privacy and operational efficiency, especially in modulus reduction operations that are not arithmetic in nature.

Innovation Solution

The method determines an approximate polynomial using Chebyshev polynomials for modulus reduction, optimizing the polynomial degree and sample points to minimize errors through L2-norm minimization, allowing for efficient homomorphic evaluation and reduced level loss in bootstrapping.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional bootstrapping algorithms are used for homomorphic encryption, then security is maintained, but computational overhead and runtime are high

Engineering Contradiction:
ImprovesecurityVSAvoidcomputational efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent changes the parameters of the polynomial approximation by optimizing the degree of the polynomial and the sampling points. This allows achieving the same security level with reduced computational complexity, thereby improving runtime efficiency while maintaining security requirements for bootstrapping in homomorphic encryption

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces traditional mechanical/computational bootstrapping algorithms with a mathematical approximation approach using polynomials. By substituting the complex traditional bootstrapping mechanism with a polynomial-based approximation system, the patent achieves faster computation while preserving the essential security properties

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If polynomial degree is increased to reduce approximation error, then accuracy improves, but computational complexity increases

Engineering Contradiction:
Improveapproximation accuracyVSAvoidpolynomial complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent optimizes the polynomial degree parameter to find the optimal balance between approximation accuracy and computational complexity. By carefully selecting the polynomial degree and sampling points, the patent achieves sufficient accuracy without unnecessarily increasing polynomial complexity, thus resolving the trade-off between precision and complexity

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If more samples are used for polynomial approximation, then approximation accuracy improves, but processing time increases

Engineering Contradiction:
Improveapproximation accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent optimizes the sampling parameters including the number of samples, their distribution, and locations. By carefully tuning these sampling parameters, the patent achieves high approximation accuracy with a reduced number of samples, thereby minimizing processing time while maintaining the required precision for secure bootstrapping

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3896895B1Method and apparatus for processing ciphertext based on homomorphic encryption
Publication Date: 2023.05.03 SAMSUNG ELECTRONICS CO LTD
  • EP3896895B1 patent drawingFigure 1
  • EP3896895B1 patent drawingFigure 2
  • EP3896895B1 patent drawingFigure 3

AI summary

A method and apparatus for processing a ciphertext based on homomorphic encryption. The method includes determining an approximate polynomial corresponding to a modulus reduction for bootstrapping a ciphertext based on samples extracted from the modulus reduction, and bootstrapping the ciphertext based on the approximate polynomial.