3
$\begingroup$

The quantum state fidelity between a pure quantum state $\rho:= \vert \psi \rangle \langle \psi \vert$ and a state $\sigma$ is \begin{align} F(\rho, \sigma):= {\rm Tr}[\sqrt{\sqrt{\rho}\sigma\sqrt{\rho}}]^2 = \langle \psi \vert \sigma\vert \psi \rangle. \end{align} Is $\langle \psi \vert \sigma\vert \psi \rangle$ equal to $\Vert \vert \psi \rangle \langle \psi \vert \sigma \Vert_1$, where $\Vert \cdot \Vert_1$ is the trace norm?

$\endgroup$

1 Answer 1

1
$\begingroup$

No, in general $\langle\psi|\sigma|\psi\rangle\ne\||\psi\rangle\langle\psi|\sigma\|_1$.

Singular value decomposition

Recall that if \begin{equation} A=\sum_i s_i|a_i\rangle\langle b_i|\tag1 \end{equation} is the singular value decomposition of operator $A$, then $\|A\|_1=\sum_i s_i$ where the positive real numbers $s_i$ are the singular values of $A$. In particular, $\||a\rangle\langle b|\|_1=1$ for every two normalized state vectors $|a\rangle$ and $|b\rangle$.

Counterexample

Take $|\psi\rangle=|0\rangle$ and $\sigma=|+\rangle\langle+|$. Then, $\langle 0|\sigma=\frac{1}{\sqrt2}\langle+|$, so \begin{align} \||\psi\rangle\langle\psi|\sigma\|_1&=\frac{1}{\sqrt2}\||0\rangle\langle +|\|_1=\frac{1}{\sqrt2}\ne\frac12=\langle\psi|\sigma|\psi\rangle.\tag2 \end{align}

Inequality

Nevertheless, we can prove that \begin{equation} \||\psi\rangle\langle\psi|\sigma\|_1\geqslant\langle\psi|\sigma|\psi\rangle.\tag3 \end{equation}

If $\sigma|\psi\rangle=0$, then the inequality is satisfied. Assume $\sigma|\psi\rangle\ne 0$ and let $s>0$ to be the L2 norm of $\sigma|\psi\rangle$. Also, define $|\phi\rangle:=\frac{\sigma|\psi\rangle}{s}$, so that $\sigma|\psi\rangle=s\cdot|\phi\rangle$. We have \begin{align} \||\psi\rangle\langle\psi|\sigma\|_1&=s\cdot\||\psi\rangle\langle\phi|\|_1\tag4\\ &=s\tag5\\ &\geqslant s\cdot\langle\psi|\phi\rangle\tag6\\ &=\langle\psi|\sigma|\psi\rangle\tag7. \end{align} This gives some meaning to the two quantities and sheds some light on why they aren't generally equal. Namely, $\||\psi\rangle\langle\psi|\sigma\|_1$ is just the L2 norm of $\sigma|\psi\rangle$, but the fidelity also depends on the angle between $|\psi\rangle$ and $\sigma|\psi\rangle$.

$\endgroup$
1
  • $\begingroup$ Thanks for correcting me! $\endgroup$ Sep 10 at 10:37

Your Answer

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

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