3
$\begingroup$

Exercise 2.63 of Nielsen & Chuang asks one to show that if a measurement is described by measurement operators $M_m$, there exists unitary $U_m$ such that $M_m = U_m \sqrt{E_m}$ where $E_m$ are the POVM associated to the measurement (that is, $E_m = M^{\dagger}_m M_m$).

I can see that, if $\sqrt{E_m}$ is invertible, then $U_m = M_m \sqrt{E_m}^{-1}$ is unitary; indeed, we have (dropping the needless subscript for simplicity) $U^{\dagger} U = (\sqrt{E}^{-1})^\dagger M^\dagger M \sqrt{E}^{-1} = (\sqrt{E}^{-1})^\dagger E \sqrt{E}^{-1} = (\sqrt{E}^{-1})^\dagger \sqrt{E} = (\sqrt{E}^{-1})^\dagger \sqrt{E}^\dagger = (\sqrt{E} \sqrt{E}^{-1})^\dagger = I^\dagger = I$ where I used that $\sqrt{E}$ is Hermitian (since it is positive).

But what if it's not invertible? Perhaps some continuity argument would work?

$\endgroup$

1 Answer 1

2
$\begingroup$

It is just a polar decomposition of $M_m$.
If $M_m = U P$ then $M^{\dagger}_m M_m = P^2$, hence $P = \sqrt{E_m}$.

Limiting argument, similar to this, also can work.

$\endgroup$

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.