Comments for Nathaniel Johnston

Comment on LaTeX Poster Template by Alex Szatmary

Alex Szatmary — Wed, 22 Feb 2012 23:27:59 +0000

I’m finishing my second poster made using this template. It’s the best thing for the job. Thanks!

Comment on LaTeX Poster Template by Nishan Mudalige

Nishan Mudalige — Mon, 13 Feb 2012 04:42:32 +0000

@Kate

To add a graphic to the top RIGHT corner, try the following code

\title{ \parbox[c]{38in}{Title_of_Poster}
\parbox[c]{6in{\includegraphics[width=5in]{graphic_name}} }

You can play around with the widths of the parboxes until you get the right fit.

Hope this helps.

Regards,

-Nishan

Comment on Counting and Solving Final Fantasy XIII-2′s Clock Puzzles by Robert

Robert — Fri, 10 Feb 2012 05:22:00 +0000

I was searching the internet thinking about the underlying structure of the clock puzzle, and I found your write up. You did a fantastic job on it; thanks for taking the time to write it!

Comment on MATLAB Scripts for Computing Completely Bounded Norms via Semidefinite Programming by Nathaniel

Nathaniel — Wed, 08 Feb 2012 22:23:14 +0000

@Ian H. – Thanks for mentioning CVX, as I was unaware of it. Getting SeDuMi to work properly is generally a pain (as you have experienced), so it’ll be nice to have an alternative to play with.

And that’s indeed some impressive speed! It could be that SDPT3 (assuming you’re using CVX’s default solver) is just faster for these types of semidefinite programs than SeDuMi is, or it’s also entirely possible that your code is just much cleaner than mine

Comment on LaTeX CV Template by Saurabh

Saurabh — Wed, 08 Feb 2012 22:21:29 +0000

Thanks a lot for the template. It is really well written and neat looking.

Comment on MATLAB Scripts for Computing Completely Bounded Norms via Semidefinite Programming by Ian H.

Ian H. — Wed, 08 Feb 2012 19:57:25 +0000

Thanks for posting this, I have found it very useful. I wrote my own diamond norm function using the cvx library for Matlab. SeDuMi is a bit problematic for me (64-bit linux, Matlab 2011b), the only version mex didn’t complain about was the newest one, which prevents me from using Kraus operators with complex entries under your code. However, I did compare the results of my code with yours in the real case, and they agree to about 9 decimal places, which is pretty good.

CVX seems to be a fair bit quicker for this application, which surprised me considering how much mex-ing SeDuMi does. I can get the diamond norm of a 4-qubit superoperator with 8 random left and right kraus operators in 16 seconds, for example.