Score:1

How to prove the inequalities of q-ary lattice determinant?

in flag

for $A\in{Z_q^{n*m}}$ and $A^{'}\in{Z_q^{m*n}}$,we have

  • $det{({\land}_q^{\bot}(A))}{\le}q^n$ and $det{({\land}_q(A^{'}))}{\ge}q^{m-n}$
  • if q is prime,and A,A' are non-singular in the finite field $Z_q$,the above inequality are equalities.

where ${\land}_q^{\bot}(A) = \{x{\in}Z^m|Ax=0{\bmod}q\}$ and ${\land}_q(A)=\{y{\in}Z^m|y=As{\bmod}q\}$

The above content comes from D. Dadush's lecture note(lemma 4) lecture_9.I don't know how to prove the above lemma because the proof in the lecture note is too sketchy for me.I would appreciate it if someone could provide more detailed proof。

LeoDucas avatar
gd flag
Maybe it is a good place to acknowledge that there are several mistakes in those lecture notes. We are currently working on fixing them :/
LeoDucas avatar
gd flag
For that particular instance, no mistakes, but maybe an explicit invocation of Lemma 10 of Lecture 2 would help https://homepages.cwi.nl/~dadush/teaching/lattices-2018/notes/lecture-2.pdf
Score:2
in flag

I get inspiration from vadim's ppt.But I only prove half of the theorem.

Proof: ${\because}{\land_q^{\bot}}(A)$ is a integer lattice, ${\therefore} {det({\land_q^{\bot}}(A))}=|{Z^m}/{\land_q^{\bot}}(A)|$. let us define a mapping ${f}:{Z^m}{\to}{Z_q^{n}}$,${f:Ax{\bmod}q}$.It's easy to verify that $f$ is homomorphic.According to the basic theorem of homomorphism,$|{Z^m}/kerf|=|{Z^m}/{\land_q^{\bot}}(A)|=|im({Z^m})|$.Because $im({Z^m}){\subseteq}{Z_q^{n}}$,so ${det({\land_q^{\bot}}(A))}=|{Z^m}/{\land_q^{\bot}}(A)|{\leq}q^n$. If q is prime,and A is non-singular,then $f$ is full homomorphism because every image in ${Z_q^{n}}$ can find the original image in ${Z^m}$.So $im({Z^m})={Z_q^{m}}$ and ${det({\land_q^{\bot}}(A))}=|{Z^m}/{\land_q^{\bot}}(A)|=q^n$.

However,I haven't found a way to prove the other half of the theorem. I would be grateful if someone could give the rest of the proof.

Ievgeni avatar
cn flag
Could you give a source about the definition of a determinant?
Mark avatar
ng flag
The second half of the theorem follows from basic lattice facts. For any lattice $\Lambda$, we have that $\det \Lambda \det \Lambda^* = 1$, where $\Lambda^*$ is the dual lattice. $\Lambda_q^\perp(A)$ and $\Lambda_q(A)$ are not dual, but they are scaled duals --- in particular $\Lambda_q(A)^* = (1/q)\Lambda_q^\perp(A)$ (and $\Lambda_q^\perp(A)^* = (1/q)\Lambda_q(A)$ ). It follows that $\det \Lambda_q^\perp(A) \det (1/q)\Lambda_q(A) = 1$, and therefore $1/\det((1/q)\Lambda_q(A)) \leq q^n$. The result then follows from simplifying.
in flag
@Mark,Thank you for your supplement. I know how to prove it.
in flag
@levgeni: I think there's something you want in this [lecture note](https://homepages.cwi.nl/~dadush/teaching/lattices-2018/notes/lecture-2.pdf)
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.