Hacker News Reader

Comment by hugh | Hacker News Reader

The paper is unconvincing. Sure, there's some mathematical similarity between Pagerank and the Schroedinger equation (they're both just eigenvector problems anyway, as I understand it) but they haven't really demonstrated that bringing in the QM formalism actually assists with calculating anything.

The "X could revolutionize Y" headline is even less convincing.

View on HN

But the QM formalism is so pretty! It MUST be useful.

'Revolutionize" may not be appropriate, but I still think the paper is a good post. For anyone with a math/physics background, this is the perfect 3min. introduction to the page-rank algorithm.

You think? I'd have thought that anybody with a good math background would better served by something like the "$25B Eigenvector" article focusing on the Linear Algebra aspects of PageRank (http://www.rose-hulman.edu/~bryan/googleFinalVersionFixed.pd...)

... and the $25B is a good 6+ minute introduction to the page-rank algorithm. The difference between the two perspectives is that the $25B paper treats PageRank as a general eigenvalue problem, while this one treats it as a very specific eigenvalue problem that has been analyzed to death by every student of QM or numerical analysis... including the associated perturbation theory, stability, preconditioning techniques, etc. As a result, the paper is very easy to scan. It is not a magnum opus, but it is cute.

Yes, that $25B eigenvector article was the one I read, and having read it, was why I didn't feel like I was getting much new from this paper.

I may have been too harsh on the paper, though, because of the headline. It's still kinda neat, just almost-certainly useless.

hugh, could you elaborate in slightly more laymen's terms? Or point to a good resource to read up on this?

It sounds interesting but is a bit over my head. Thanks!

Reply on news.ycombinator.com