0
$\begingroup$

I'm having a bit of trouble understand @DaftWullie's answer here.

I understood that the $4\times 4$ matrix $A$ $$ \frac{1}{4} \left[\begin{matrix} 15 & 9 & 5 & -3 \\ 9 & 15 & 3 & -5 \\ 5 & 3 & 15 & -9 \\ -3 & -5 & -9 & 15 \end{matrix}\right]$$

can be decomposed into Pauli matrices as:

$$A=15\mathbb{I}\otimes\mathbb{I}+9Z\otimes X+5X\otimes Z-3Y\otimes Y$$

So far so good.

Then he says:

Now, it is interesting to note that every one of these terms commutes. So, that means that $$ e^{iA\theta}=e^{15i\theta}e^{9i\theta Z\otimes X}e^{5i\theta X\otimes Z}e^{-3i\theta Y\otimes Y}. $$ You could work out how to simulate each of these steps individually, but let me make one further observation first: these commuting terms are the stabilizers of the 2-qubit cluster state. That may or may not mean anything to you, but it tells me that a smart thing to do is apply a controlled-phase gate. $$ CP\cdot A\cdot CP=15\mathbb{I}\otimes\mathbb{I}+9\mathbb{I}\otimes X+5X\otimes \mathbb{I}-3X\otimes X. $$

Now, I hadn't heard of clusters states before, but Wikipedia gave me some idea (I'll probably need to through it a few more times though).

Anyhow, as far as I know, the controlled phase gate $CP$ is basically a controlled-$R_{\phi}$ gate where $R_{\phi}$ is:

$$\left[\begin{matrix}1 & 0 \\ 0 & e^{i\phi}\end{matrix}\right]$$

So, controlled $R_{\phi}$ would be

$$\left[\begin{matrix} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & e^{i\phi} \end{matrix}\right]$$

This is where I am confused. Shouldn't $(\text{controlled }R_{\phi})A(\text{controlled }R_{\phi})$ contain a $\phi$ term somewhere? I don't understand how its Pauli decomposition $15\mathbb{I}\otimes\mathbb{I}+9\mathbb{I}\otimes X+5X\otimes \mathbb{I}-3X\otimes X$ contains no term containing the phase angle $\phi$. Wolfram Alpha also agrees that the matrix multiplication result of $\operatorname{CP\cdot A\cdot CP}$ must contain a phase term. So, I'm not quite sure how Pauli decomposition of $\operatorname{CP\cdot A\cdot CP}$ as stated by DaftWullie in his answer arises. Am I missing something?

$\endgroup$
1
$\begingroup$

When I talked about a controlled-phase gate, I meant the standard gate that has unitary matrix $$ \left(\begin{array}{cccc} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & -1 \end{array}\right) $$

Note that this is related to the controlled-not via the action of a Hadamard on the target qubit. It also satisfies some standard propagation relations on Pauli operations: $$ Z_n\mapsto Z_n\qquad X_n\mapsto Z_{3-n}X_n $$ for $n\in\{1,2\}$ being the two qubits that the controlled-phase gate acts on. enter image description here

$\endgroup$
  • $\begingroup$ Isn't this normally called the controlled-Z gate? It is a special case of the controlled phase shift gates when $\phi = \pi$ as far as I know. Also, if I consider $\phi$ to be $\pi$, then according to Wolfram $CP\cdot A\cdot CP$ should be $\text{diag}(15/4,15/4,15/4,-15/4)$. And for that, this Python script gives me $1.875 (\Bbb I \otimes \Bbb I ) + 1.875( \Bbb I \otimes \sigma_z )+1.875( \sigma_z \otimes I ) -1.875( \sigma_z \otimes \sigma_z )$. $\endgroup$ – Sanchayan Dutta Oct 19 '18 at 12:24
  • $\begingroup$ Maybe Wolfram is doing the matrix multiplication wrong. I'm checking again $\endgroup$ – Sanchayan Dutta Oct 19 '18 at 12:25
  • $\begingroup$ @Blue In terms of terminology, I guess you can use controlled-phase to mean arbitrary phase, but it is also used extensively meaning specifically $\phi=\pi$. The top google scholar results for the term all return $\phi=\pi$ instances. $\endgroup$ – DaftWullie Oct 19 '18 at 12:37
  • $\begingroup$ Thanks, makes sense! I realized that using the term matrix multiply in Wolfram Alpha causes it to multiply the matrices "element-wise", which is why it was giving the incorrect result. This time it looks OK: $$\left[\begin{matrix}15 & 9 & 5 & 3 \\ 9 & 15 & 3 & 5 \\ 5 & 3 & 15 & 9 \\ 3 & 5 & 9 & 15\end{matrix}\right]$$ which decomposes into $$15(I \otimes I)+9( I \otimes X)+5(X \otimes I )+3(X \otimes X ) $$ according to the script. So apparently there was a sign error in the last term of Pauli decomposition in the other answer (which you had mentioned the possibility of). $\endgroup$ – Sanchayan Dutta Oct 19 '18 at 12:55
  • $\begingroup$ This is the link for the matrix multiplication $\text{CP}\cdot A \cdot \text{CP}$. So, if I'm not mistaken, I just need to change $R_x(-3\theta)$ to $R_x(3\theta)$ in the final circuit to correct it, right? $\endgroup$ – Sanchayan Dutta Oct 19 '18 at 12:58
1
$\begingroup$

Yes if you work with general phase shift there would be $\phi$ in the final answer. In fact you would be able to take $\phi=0$ and just get $A$ back. Try $\phi=\frac{\pi}{4}$. Looks like notational mismatch of what's called a phase gate/phase shift gate. Whether it means the entire 1 parameter family or just one specific value of $\phi$.

EDIT: Incorrect, see comment below.

$\endgroup$
  • $\begingroup$ Should have checked beforehand instead of guessing the value that is often called phase. $\endgroup$ – AHusain Oct 19 '18 at 12:31

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.