In celebration of the recent resolution of the Kadison-Singer problem by Marcus, Spielman, and Srivastava, here is a question on isotropic point sets on which Kadison-Singer does not (seem to) shed any light. A positive resolution would likely have strong implications for the Sparsest Cut problem and SDP hierarchies. The question arose in discussions with Shayan Oveis Gharan, Prasad Raghavendra, and David Steurer.

**Open Question:** Do there exist constants such that for any , the following holds? Let be a collection of orthonormal bases and define . Then there are subsets with and .

[Some additional notes: One piece of intuition for why the question should have a positive resolution is that these orthonormal bases which together comprise at most vectors cannot possibly “fill” -dimensional space in a way that achieves -dimensional isoperimetry. One would seem to need points for this.

One can state an equivalent question in terms of vertex expansion. Say that a graph on vertices is a *vertex expander* if for all subsets with . Here, denotes all the nodes that are in or are adjacent to . Then one can ask whether there exists a 1-1 mapping from to for some orthonormal bases such that the endpoints of every edge are mapped at most apart (as ).]

What would be the implication for sparsest cut?

Comment by minilek — June 27, 2013 @ 5:22 pm

Strictly speaking, there is no direct implication because this question is the “hard part” of a more general question. The more general question is to find such separated sets given any set of unit vectors in isotropic position.

The implication is as follows: For every , there is a number such that Sparsest Cut admits a -approximation in time , i.e. sub-exponential time constant factor approximations to (uniform) Sparsest Cut.

Comment by James L. — June 27, 2013 @ 5:26 pm

What is your precise definition of isotropic position? Is it that the covariance matrix (the sum of uu^T over vectors u in the set) is a multiple of the identity or is it exactly the identity? Seems like it’s the former, but just want to make sure.

Comment by NG — June 28, 2013 @ 8:25 pm

More specifically, the general condition is as follows: Given unit vectors , we assume that . Of course, a special case is when the vectors can be partitioned into orthonormal bases.

Comment by James L. — June 28, 2013 @ 10:25 pm

Is it important that the bases are over the reals? I think I can see a construction of such sets over complex vector spaces.

Comment by sflammia — July 5, 2013 @ 1:06 am

A general construction of such A,B over would be equally remarkable (and useful).

Comment by James Lee — July 5, 2013 @ 11:59 am

Wow, awesome blog layout! How long have you been blogging

for? you made blogging look easy. The overall look of your web site is wonderful, as well as the

content!

Comment by reading glasses with built in lights — December 6, 2013 @ 4:33 am