Encrypted Computation Parameter Tuning Under Ciphertext Noise Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing homomorphic encryption schemes face challenges in selecting encrypted computation parameters that balance security, accuracy, and computational efficiency, particularly due to the accumulation of noise during multiple operations, which can lead to decryption failures.
Innovation Solution
A method is developed to optimize encrypted computation parameters by dividing the computation graph into subgraphs based on recurring patterns, using noise constraint functions and cost functions to minimize computational cost while ensuring security and accuracy, and determining optimal parameters for key switching and bootstrapping keys.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If noise is increased to make attacks harder, then security is improved, but decryption accuracy deteriorates
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the noise level in ciphertexts based on the specific security requirements and computation depth. Different noise parameters are selected for different encryption operations, allowing the system to maintain sufficient security while preserving decryption accuracy. The optimization process determines optimal noise parameters that balance security hardness with computational correctness.
2Adaptability or versatility
If multiple homomorphic operations are performed, then computational functionality is improved, but noise accumulation increases leading to decryption failure
Solution Approach 1:
The patent applies preliminary action by performing noise management operations (such as bootstrapping or noise reduction) before the noise level becomes critical. The system predicts noise accumulation based on the computation graph and schedules noise reduction operations in advance, preventing decryption failure before it occurs. This allows deeper computational functionality while maintaining reliability.
Solution Approach 2:
The patent implements feedback mechanisms where the noise level is continuously monitored and measured during the homomorphic computation process. Based on the measured noise levels, the system dynamically adjusts subsequent operations or triggers noise reduction operations. This closed-loop control ensures that noise remains within acceptable bounds while maximizing computational functionality.
3Measurement precision
If bootstrapping is performed to reduce noise, then decryption accuracy is improved, but computational cost increases
Solution Approach 1:
The patent applies local quality by selectively applying bootstrapping operations only to specific ciphertexts or specific points in the computation graph where noise has accumulated to problematic levels. Instead of uniformly applying bootstrapping to all operations, the system identifies critical locations where noise reduction is most beneficial, thereby improving decryption accuracy where needed while minimizing the overall computational overhead.
4Reliability
If encryption parameters are optimized for security, then security is improved, but computational efficiency deteriorates
Solution Approach 1:
The patent applies dynamics by making encryption parameters adaptive rather than static. The system dynamically selects encryption parameters (such as noise levels, modulus sizes, and bootstrapping configurations) based on the specific computation being performed, the security requirements, and the available computational resources. This allows the system to achieve high security when needed while maintaining computational efficiency for less sensitive operations.
Data Source
Figure 1A~1C
Figure 2
Figure 3
AI summary
Some embodiments are directed to a computer-implemented method of determining encrypted computation parameters for carrying out an encrypted computation on noisy ciphertexts. A computation graph is divided into multiple subgraphs, defined by a type and by instantiation parameters for the type. Respective sets of encrypted computation parameters are defined for the respective types. An optimization of the encrypted computation parameters is performed to minimize a computational cost of carrying out the encrypted computation according to the encrypted computation parameters. The encrypted computation parameters are constrained to satisfy a noise constraint on ciphertext noise while carrying out the encrypted computation. The noise constraint is based on respective noise constraints for respective subgraphs, defined by a noise constraint function for the type that takes at least the encrypted computation parameters for the type and the instantiation parameters of the subgraph as input.