SIGACT CG Column: Moving heaven and earth

My new column for SIGACT News is up (non pay-walled version here). In it, I discuss two different ways of defining distances between distributions.

The earthmover distance is a familar measure to TCS folk, and the kernel distance (or the maximum mean distortion, or the distance covariance) is a familiar character in machine learning and statistics.

It turns out that while statistically, these are not as different as one might think, they are different when it comes to computing them, and understanding their geometry.

The column describes both measures and their properties, and does a compare-contrast between them. More generally, I try to describe how one should think about the process of constructing a distance function between two distributions.

The SIGACT CG column

As you might have just discovered, I'm the second half of the two-headed monster that's taken over the SIGACT Geometry Column after Joe O'Rourke stepped down (Adrian Dumitrescu, who hopefully does not mind being referred to as the head of a monster, is the other half). My first column is up, and it talks about some recent fascinating developments in nonnegative matrix factorization.

My hope with the pieces I write is to cover areas of geometry that may have not had sufficient representation in the past (especially things closer to problems I'm interested in). My next column is due in August, and apart from doing a wrapup on SoCG, other things that come to mind include Laplacians and graph geometry, reproducing kernels, or even Bregman geometry.

But everything is now mobile, crowdsourced and social networked, so I'm looking for your suggestions on interesting topics to cover, new emerging areas, results that I'm not tracking, and so on. So post away here or on G+.