2
$\begingroup$

A non-entangling gate on $ n $ qubits is an element of the group $$ N\Big(\bigotimes_{i=1}^n U(2)\Big)=\bigotimes_{i=1}^n U(2) \rtimes S_n $$ which is generated by $ U(2) $ acting locally on each qubit as well as all $ SWAP $ gates between qubits.

Let $ C_1,C_2 $ be $ [[n,k,d]] $ quantum codes. In other words, $ n_1=n_2, k_1=k_2,d_1=d_2 $. If $ C_1, C_2 $ are equivalent by non-entangling gates then they have the same weight enumerator.** Is the converse true? That is, if $ C_1,C_2 $ have the same weight enumerator must they be related by a non-entangling gate?


Note that the weight enumerator already gives $ d $ (it is the least degree of a nonzero term in $ B(x)-A(x) $ enumerators polynomials and distance for a quantum error correcting code, note that $ A(x) $ determines $ B(x) $ by quantum McWilliams identity) and gives $ n-k $ (because $ 2^{n-k}=A(1) $ Do the coefficients of the weight enumerator polynomial add up to $2^{n-k}$ for any $[\![n,k]\!]$ code?) so two different codes on $ n $ qubits with the same weight enumerator must have the same parameters $ [[n,k,d]] $.


**For a reference for this fact see top of page 4 in

Quantum Weight Enumerators

which in turn refers to the discussion below equation (9) in

Quantum MacWilliams Identities

$\endgroup$

1 Answer 1

2
$\begingroup$

No. There are 16 inequivalent $[[7,1,3]]$ codes but they correspond to only 10 weight enumerators. For reference see https://arxiv.org/pdf/0709.1780.pdf.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.