Thank you for all of you work and for this project .
I am trying to use binary gates in a sequence (approx ~70) with the best performance. But I see that there are some differences when decrypted compared to the operations done on the clear ciphertext. I have a median similarity of 70% (i.e. 70% of the decrypted cipher text is the same as the clear file).
So firstly, I would like to know if my use of the binary gates was wrong ?
Also, I would like to know if there has been some verifications/statistics about noise propagation when boostrapping ? From what I understood from the code, after each gate there is bootstrapping.
I copy pasted my code down there if you want to have a look, this one gives ~52% accuracy
Dear all,
Thank you for all of you work and for this project .
I am trying to use binary gates in a sequence (approx ~70) with the best performance. But I see that there are some differences when decrypted compared to the operations done on the clear ciphertext. I have a median similarity of 70% (i.e. 70% of the decrypted cipher text is the same as the clear file).
So firstly, I would like to know if my use of the binary gates was wrong ?
Also, I would like to know if there has been some verifications/statistics about noise propagation when boostrapping ? From what I understood from the code, after each gate there is bootstrapping.
I copy pasted my code down there if you want to have a look, this one gives ~52% accuracy
Thank you !
The function that does the same equation but on the clear array: