High density SIS and Low density SIS

Lee Seungwoo

1/2/24, 2:31 AM

I am searching for the exact definition of High density SIS and Low density SIS, but there is something unclear about it.

SIS problem is to find $x\in \mathbb{Z}^m$ such that for random $A\in\mathbb{Z}_q^{n\times m}$, $Ax=0$ and $\lVert x\rVert < \beta$.

As far as I know, high density SIS is when $n\log q < m$, so that the number of inputs are larger than the output and hence there are many solutions.

My question is :

Is low density SIS parameterized by $n\log q > m $?
If so, are there any specific situations that for some situations like, high density SIS must be used rather than low density one.. and for the others, low density SIS must be used.. ?

Thank you in advance.

0 + 2

lattice-crypto

sis

Geoffroy Couteau

1/2/24, 9:02 AM

I am not an expert (at all) in SIS, but a slightly educated guess would be: if you want to build a collision-resistant hash function, you will usually be in a setting where there are multiple solutions (since there are many preimages to a given hash). The injective setting is probably more common in encryption-style applications.

Lee Seungwoo

1/4/24, 1:17 AM

@GeoffroyCouteau That would be an answer to my question 2. Thank you for the answer. Have a nice day!

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: High density SIS and Low density SIS

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.