Jean-Philippe Bossuat

Results 36 comments of Jean-Philippe Bossuat

The issue you are running into is not related to the relinearization key but to the homomorphic capacity of the ciphertext. The error says that you would need 5 levels...

That's because the two schemes are fundamentally identical, they only differ by their encoding and tensoring operation. So until you do a multiplication, the noise budget between the two schemes...

The same method is used for both schemes: [Decryptor::invariant_noise_budget](https://github.com/microsoft/SEAL/blob/main/native/src/seal/decryptor.cpp): `log(norm(dec(ct))) - log(norm(plaintext)) - 1 = log(norm(err)/2)`.

A ciphertext that encrypts a message of `n` slots in a ring degree of `2^{k+1}n` is equivalent to a ciphertext that encrypts `2^{k}` concatenations of this message but encoded as...

Yes my bad, I should have added a code snippet. Here is the bootstrapping example that triggers the bug ```Go // Package main implements an example showcasing the basics of...

The method `AtLevel` is a convenience method that is supposed to return an instance of the same sampler, but that will sample at a given level. It is expected that...

Great initiative :) Some precision about the unification of BFV/BGV, it is unified under the same scaling factor: T^-1 mod Q, i.e. the BFV scaling factor.

`EqualSlice` is not constant time anymore after these changes

> > `EqualSlice` is not constant time anymore after these changes > > Ah yes, but before it would also panic early depending on the size of `b`. > >...