Skip to main content
9 events
when toggle format what by license comment
May 6, 2021 at 22:27 history edited Cosmas Zachos CC BY-SA 4.0
added 694 characters in body
May 6, 2021 at 22:05 comment added Cosmas Zachos I'm sorry, The Pauli vector method is also overkill. Will write a simpler answer. Note N can be transformed to a symmetric, hence orthogonally diagonalizable matrix by a further similarity transformation...
May 6, 2021 at 21:52 comment added a_guest It appears I was too fixated on arriving at that $\exp$ term, so I got the order wrong. Of course it should be $N^n = \exp n\log N$. But here I don't see how this can be converted to the form $\exp ia(\hat{z}\vec{\sigma})$. How do I determine $a$ in this case?
May 6, 2021 at 21:34 comment added a_guest So in the end I could express $M^n$ as a function of $\mathbb{1}$ and $\hat{z}\vec{\sigma}$ by using the binomial theorem? In order to use Euler's formula, I would start by writing $M^n = \log\exp nM$ and then transform this expression until I arrive at the form $\exp ia(\hat{z}\vec{\sigma})$ (with $a = -inz$). Is that correct?
May 6, 2021 at 21:26 vote accept a_guest
May 6, 2021 at 15:49 comment added Brick As an add-on: This procedure could be viewed as a specific case of normalizing $M$, which is useful and, if you're computing numerically, sometimes necessary for the results to make sense or to be numerically stable. Even if you didn't have the problem with units, if the singular values are very different in magnitude you might still have problems that could be alleviated by a modified application of this procedure. This is especially true if your matrix is related to analysis of random variables, e.g. stats.stackexchange.com/questions/12200/…
May 6, 2021 at 15:33 history edited Cosmas Zachos CC BY-SA 4.0
added 50 characters in body
May 6, 2021 at 13:36 history edited Cosmas Zachos CC BY-SA 4.0
added 176 characters in body
May 6, 2021 at 13:27 history answered Cosmas Zachos CC BY-SA 4.0