Score:0

How to understand noise growth in BFV?

es flag

I am trying to understand the noise growth due to multiplication in BFV encryption.

As explained in section 4 and equation 3 of this paper: https://eprint.iacr.org/2012/144.pdf.

I couldn't follow what is $r_a$ and $r_r$ is in their equations.

Also how they are bounding the values of all rounding errors?

Also is there are simpler explanation?

Score:1
in flag

$\boldsymbol{r}_a$ is the approximation error for Equation 3, since we are creating an approximation where we round the ciphertext values.

$\boldsymbol{r}_r$ are the non-integer values in the equality generated by scaling Equation 2 by $t/q$, i.e.: $$\textbf{r}_r=\frac{t}{q}\cdot[\textbf{v}_1\cdot\textbf{v}_2]_\Delta-\frac{r_t(q)}{q}\cdot (\Delta\cdot \textbf{m}_1\cdot\textbf{m}_2+(\textbf{m}_1\cdot\textbf{v}_2+\textbf{m}_2\cdot\textbf{v}_1)+\textbf{r}_v$$ Note that these will be the only values affected by the rounding. The error from this rounding is bounded by $\lVert r_r\rVert$ (since $[\boldsymbol{r}_r]_q=\sum_{i=0}{[r_{r_i}]_q\cdot x^i})$

As for a simpler explanation I've yet to find one, if you only need the implementation and not the correctness perhaps consider this site.

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.