Approximating Max Function in Fully Homomorphic Encryption

This study focuses on efficiently finding the location of maximum value for large-scale values encrypted by CKKS (Cheon—Kim—Kim–Song) method. To find value, logM+1 comparison operations and logM rotation operations, 2logM+3 additions 2logM+1 multiplications are required. However, there is no known way to a k-approximate i.e., with same most significant k-bits as value. In this study, when range all data in each slot ciphertext [0, 1], we propose method positions whose match The proposed can slots from input ciphertexts where their have performing 2k (4k+2) multiplications, (6k+2klogM+3) additions, 2klogM operations. Through experiments complexity analysis, show that more efficient than existing locations k MSB equal result be applied various privacy-preserving applications environments, such IoT devices.