PSD lifting and Unique Games integrality gaps

By now, it is known that integrality gaps for the standard Unique Games SDP (see the paper of Khot and Vishnoi or Section 5.2 of this post) can be used to obtain integrality gaps for many other optimization problems, and often for very strong SDPs coming from various methods of SDP tightening; see, for instance, the paper of Raghavendra and Steurer.

Problematically, the Khot-Vishnoi gap is rather inefficient: To achieve the optimal gap for Unique Games with alphabet size {L}, one needs an instance of size {\exp(\Omega(L))}. As far as I know, there is no obstacle to achieving a gap instance where the number of variables is only {\mathrm{poly}(L)}.

The Walsh-Hadamard code

The Khot-Vishnoi construction is based on the Hadamard code.
(See Section 5.2 here for a complete description.) If we use {L^2(\{-1,1\}^k)} to denote the Hilbert space of real-valued functions {f : \{-1,1\}^k \rightarrow \mathbb R}, then the Walsh-Hadamard basis of {L^2(\{-1,1\}^k))} is the set of functions of the form

\displaystyle  u_S(x) = \prod_{i \in S} x_i,

where {S \subseteq \{1,2,\ldots,k\}}.

Of course, for two such sets {S \neq T}, we have the orthogonality relations,

\displaystyle  \langle u_S, u_T \rangle = 0.

In their construction, the variables are essentially all functions of the form {f : \{-1,1\}^k \rightarrow \{-1,1\}}, of which there are {2^{2^k}}, while there are only {2^k} basis elements {\{u_S\}_{S \subseteq [k]}} which act as the alphabet for the underlying Unique Games instance. This is what leads to the exponential relationship between the number of variables and the label size.

A PSD lifting question

In an effort to improve this dependence, one could start with a much larger set of nearly orthogonal vectors, and then somehow lift them to a higher-dimensional space where they would become orthogonal. In order for the value of the SDP not to blow up, it would be necessary that this map has some kind of Lipschitz property. We are thus led to the following (possibly naïve) question.

Let {C(d,\varepsilon)} be the smallest number such that the following holds. (Here, {S^{d-1} \subseteq \mathbb R^d} denotes the {(d-1)}-dimensional unit sphere and S(L^2) denotes the unit-sphere of L^2.)

There exists a map {F : S^{d-1} \rightarrow S(L^2)} such that {\|F\|_{\mathrm{Lip}} \leq C(d,\varepsilon)} and whenever {u,v \in \mathbb R^d} satisfy {|\langle u,v\rangle| \leq \varepsilon}, we have {\langle F(u), F(v)\rangle = 0}.

(Recall that \|F\|_{\mathrm{Lip}} = \sup_{x \neq y \in S^{d-1}} \|F(x)-F(y)\|/\|x-y\|.)

One can show that

\displaystyle C(d,\varepsilon) \lesssim \frac{\sqrt{d}}{1-\varepsilon}

by randomly partitioning {S^{d-1}} so that all vectors satisfying {|\langle u,v\rangle| \leq \varepsilon} end up in different sets of the partition, and then mapping all the points in a set to a different orthogonal vector.

My question is simply: Is a better dependence on {d} possible? Can one rule out that {C(d,\varepsilon)} could be independent of {d}? Note that any solution which randomly maps points to orthogonal vectors must incur such a blowup (this is essentially rounding the SDP to an integral solution).

3 thoughts on “PSD lifting and Unique Games integrality gaps

  1. This may be a proof that the Lipschitz constant C(d,\varepsilon) must depend on d.

    Call a function f:[-1,1]\to \mathbb R positive definite on \mathbb S^d if for every n and every Gram matrix A = (a_{i,j}) of n points on \mathbb S^d, the matrix (f(a_{i,j})) is positive semi-definite.

    A map F_d:\mathbb S^d\to \mathbb S(L^2) can be assumed to come from a positive definite function f :[-1,1]\to \mathbb R on \mathbb S^d, in the sense that \langle F(u),F(v)\rangle = f(\langle u,v \rangle) for all u,v\in\mathbb S^d. This should follow by symmetrizing \mathbb F_d by isometries of \mathbb S^d.

    Specifically, we redefine \mathbb F_d(u) by the direct integral \int^\oplus F_d(\rho u) \;d \rho, where the integral is over the orthogonal group O(d+1). I believe this step can be done, though I didn’t check the details carefully.

    Next, consider a sequence of maps f_d: [-1,1]\to \mathbb R positive definite on \mathbb S^d. Passing to a subsequence, they converge pointwise to a function f:[-1,1]\to \mathbb R positive definite on the Hilbert sphere \mathbb S^\infty. Any such function f must be analytic in the open interval (-1,1), thanks to a result of Christensen and Ressel

    The orthogonality assumption about F_d (and hence f_d) implies the limit function f vanishes identically on (-1,1). Hence the sequence of maps F_d cannot share a common Lipschitz constant independent of dimension d.

    The drawback of this proof is the lack of an effective lowerbound on the dependence.

  2. Hi, this certainly appears to be a correct proof that C(d,\varepsilon) \to \infty as d \to \infty, and the final conclusion about analycity of f follows directly from Schoenberg’s characterization. Very nice.

    I suspect that a quantitative bound can be obtained from the characterization for positive definite functions on \mathbb S^{d-1}; see here.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s