I'm working in some homomorphic encryption for my Ph.D. but I'm not of the field... So I'm possible going to ask a really simple question that I'm struggling to find an answer on the internet. Basically I'm using an scheme to encrypt some input (a vector of doubles/integers), change one bit (bit flip) to the encoding/encryption, continue the pipeline, decrypted and compare the input/output that ideally are equal. So I'm trying to figure out (and understand why) the best metrics to measure this different/error. The elements of the input vector has no relation between them, there are simple independent values that I randomly sample. For now I'm using norm2 and Hamming distance, but without a deep knowledge...
Some ideas? Some guideline?
