4
$\begingroup$

I have been reading this paper but don't yet understand how to implement a circuit to determine in which state the qubit is not for a cyclic POVM. More specifically, I want to implement a cyclic POVM with $m=3$;

Update: I came across having to implement this unitary matrix: $$ M= \frac{1}{\sqrt{2}}\left[ {\begin{array}{cc} 1 & 1 \\ 1 & w \\ \end{array} } \right] $$ Where $w$ is a third root of unity using rotations, after which I am stuck.

$\endgroup$
  • 1
    $\begingroup$ That's not a unitary matrix unless w=-1.. $\endgroup$ – Craig Gidney Mar 6 at 19:49
5
$\begingroup$

This is not the unitary that you have to implement: you need a two-qubit unitary $$ \frac{1}{\sqrt{3}}\left(\begin{array}{cccc} 1 & 1 & 1 & 0 \\ 1 & \omega & \omega^2 & 0 \\ 1 & \omega^2 & \omega & 0 \\ 0 & 0 & 0 & \sqrt{3} \end{array}\right), $$ where $\omega=e^{2i\pi/3}$, the point being that if you introduce an ancilla qubit in the state 0, apply this unitary, and then measure in the computational basis, the 3 measurement outcomes 00, 01 and 10 correspond to the 3 POVM elements.

I don't (yet) have a circuit implementation for this. You'll see the paper you cite carefully avoids talking about the Fourier transform in non-power of 2 dimensions. You certainly could use the standard constructions based on Givens rotations, but the result is going to be fairly horrible.

Here's an attempt at a circuit. I've made a few tweaks since I last ran it through a computer to check, so it's always possible that a slight error has crept in, but broadly... enter image description here Here, I'm using $Z^r$ to denote $$ \left(\begin{array}{cc} 1 & 0 \\ 0 & e^{i\pi r} \end{array}\right), $$ and $$ V=\frac{1}{\sqrt{3}}\left(\begin{array}{cc} 1 & \sqrt{2} \\ -\sqrt{2} & 1 \end{array}\right). $$

$\endgroup$
  • $\begingroup$ Hmm, I was actually talking about F2, not the whole unitary. Or, I am missing something? $\endgroup$ – chubakueno Mar 4 at 16:09
  • $\begingroup$ @chubakueno You said you wanted $m=3$, which means you need $F_3$, which is a $3\times 3$ matrix which we must embed into a $4\times 4$ matrix if we're using qubits. $\endgroup$ – DaftWullie Mar 4 at 16:11
  • $\begingroup$ I understand now. So, is there an example of IQFT for non power of twos there? I would be surprised to find there isn't, but what do I know. $\endgroup$ – chubakueno Mar 4 at 16:26
  • $\begingroup$ @chubakueno I didn't immediately find one. I've now constructed one for the $m=3$ case, but don't have time right now to write it into a circuit. $\endgroup$ – DaftWullie Mar 4 at 16:48
  • 1
    $\begingroup$ @xbk365 once i’m done with the evening’s childcare responsibilities... $\endgroup$ – DaftWullie Mar 4 at 18:21
2
$\begingroup$

Check this
$$\frac{1}{\sqrt{3}}\left(\begin{array}{ccc} 1 & 1 & 1 \\ 1 & \omega & \omega^2 \\ 1 & \omega^2 & \omega \\ \end{array}\right) = \left(\begin{array}{cc} H & 0 \\ 0 & 1 \\ \end{array}\right) \cdot \frac{1}{\sqrt{3}}\left(\begin{array}{ccc} \sqrt{2} & 0 & 1 \\ 0 & \sqrt{3} & 0 \\ 1 & 0 & -\sqrt{2} \\ \end{array}\right) \cdot M_{3} $$

$$M_{3} = \left(\begin{array}{ccc} 1 & 0 & 0 \\ 0 & \frac{1}{\sqrt{2}}i\omega^2 & \frac{1}{\sqrt{2}}i\omega \\ 0 & -\frac{1}{\sqrt{2}}\omega^2 & -\frac{1}{\sqrt{2}}\omega \\ \end{array}\right) $$

$$ \frac{1}{\sqrt{2}}\left(\begin{array}{ccc} i\omega^2 & i\omega \\ -\omega^2 & -\omega \\ \end{array}\right) = X \cdot S \cdot X \cdot Z \cdot H \cdot \left(\begin{array}{ccc} \omega^2 & 0 \\ 0 & \omega \\ \end{array}\right) \cdot Z $$

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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