## Alternating methods for SVDNovember 9, 2010

Posted by Sarah in Uncategorized.
Tags: , ,
1 comment so far

Last week’s applied math seminar was Arthur Szlam, talking about alternating methods for SVD. I’m putting up some notes here; several results are presented without proof.

Given an m x n matrix A, we want to find the best rank k approximation to A.
There is an explicit solution to this problem, known as the singular value decomposition. We write
$A = U \Sigma V^T$
where $\Sigma$ is diagonal, $U, V$ are orthogonal. The columns of $U$ are the eigenvectors of $A^T A$ and the columns of $V$ are the eigenvectors of $A A^T$. This costs $O(m^2 n)$ operations, and it’s a completely explicit solution. To get a rank-k approximation, we take
$\hat{U} \hat{\Sigma} \hat{V}^T$
where the “hat” means taking the first k columns of each matrix.
This is the best rank-k approximation, both in the sense of Frobenius norm and operator norm.

However, if $k << m$, $O(m^2 n)$ time is wasteful. There are recent attempts to do this faster, in $O(mkn)$ time.

We fix k, and a small oversampling parameter p, and let $l = k + p$. Draw an $l \times m$ random Gaussian matrix $\Omega$, and set $Q^T = \Omega A$.
Calculate the SVD, $\tilde{U} \tilde{\Sigma} \tilde{V}^T = Q$, and set $U$ to be the first k columns of $\tilde{U}$, etc.

Why does this work?

Well, pretend A is really of rank 5, for instance. Hit A with a random matrix. It is unlikely that any vector will be in the kernel of A. Whatever's left is in the range. So we're very likely to have the full range of A; we probably preserve the SVD. However, in real life we don't have a known rank k matrix.

What we have is a guarantee of quality, in terms of the decay of the singular values:
$E ||A - A Q Q^T|| \le [ 1 + \sqrt{\frac{k}{p-1}} + \frac{e \sqrt{k + p}}{p} \sqrt{m-k}] \sigma_{k +1}$

Here the expected value is taken over the random Gaussian draw, the norm is the operator norm, and $\sigma_{k + 1}$ is the k + 1th singular value; if A is really rank k, this is zero, and the expected error is therefore zero. If $\sigma_{k+1}$ is small, then the bound is good. We're in danger if m is large and $\sigma_{k+1}$ is not super small.

This actually happens, unfortunately. For example, in term document matrices, where a collection of text documents form the rows, and words form the columns, with a word count in each matrix entry indicating how often a word was used in each document. (A little imagination should indicate the importance of these kinds of matrices for semantic analysis.) These matrices are sparse and large, but their singular values decay slowly. So what can we do here?

We hit $\Omega$ by some more powers of A, to drive down the smaller eigenvalues. In practice, 10 or 15 times.
$Q_q^T = \Omega A (A^TA)^q$
Now we have a new estimate, depending on the power q.

$E ||A - A Q_q Q_q^T|| \le [1 + \sqrt{\frac{k}{p-1}} + \frac{p \sqrt{k + p}}{p} \sqrt{m-k}]^{1/2q + 1} \sigma_{k+1}$

This works, but it can be memory intensive to save all the intermediate powers. This is where the "alternating" method comes in.

Here we look at a related problem: matrix completion. You're given a subset of the entries of a matrix A, and you know A is low rank. How do you fill in the rest of the entries? The simplest approach is alternating least squares.

Write $A = PQ^T$ where P is m-by-k and Q is n-by-k.
Alternate between updates of P and Q:
$P_{j+1} = argmin ||PQ_j^T - A||^2$
$Q_{j+1} = argmin ||P_{j-1} Q^T - A||$
These are least-squares problems, and can be solved explicitly using

Step 1: form $m \times l$ Gaussian, $P_0$.
Step 2: form $n \times l$ matrix $Q_0$ st
$Q_0 R_0 = (P_0^T A)^T$
(a QR-decomposition, with Q orthogonal and R upper triangular.)
Step 3: iterate for $j = 0 \dots j = q-1$
$P_{j+1} R_{2j - 1} = A Q_j$
$Q_{j + 1} R_{2j +2} = (P_{j+1}^T A)^T$

Coming back to our own problem, we can use the same formula to calculate $Q_q^T$.
$Q_j = (A^T A)^j A^T P_o (\prod_{s = 0}^{2j} R_s^{-1}.$

This turns out to work well in practice; the example set was a database of labeled faces, which have the first 20 eigenvalues decaying fast, then the next decaying very slowly. This kind of alternating SVD works well for faces.

## Slides for Random Matrix TheoryApril 25, 2010

Posted by Sarah in Uncategorized.
Tags:

A very useful introduction to what’s going on recently in random matrix theory in Roman Vershynin’s slides here. What amazed me, seeing the big picture, was that so many of the fundamental results are quite recent. Universality of the Wigner semicircle law was only proven in 1973 and the circular law not till ’84. And everything I really want to know — what to do with sparse matrices, perturbed random matrices, Fourier matrices — is largely open.
More on what all this has to do with compressed sensing.

H/t

## Bad News for PCAApril 20, 2010

Posted by Sarah in Uncategorized.
Tags: ,

I just saw a talk by Raj Rao Nadakuditi about random matrix theory and the informational limit of eigen-analysis.

The issue is this. Lots of techniques for dealing with large data sets — PCA, SVD, signal detection — call for analyzing the eigenvectors and eigenvalues of a large matrix and reducing the dimensionality. But when do these friendly techniques fail us? It turns out that there is a phase transition, a signal-to-noise ratio below which we cannot distinguish the signal eigenvalues and eigenvectors from the noise. No matter how much data we collect, eigenvector-based dimensionality techniques will not distinguish signal from noise.

The simplest example Nadakuditi deals with is a Gaussian matrix perturbed by a rank-one matrix.
is a Gaussian random matrix, is a normalized symmetric random matrix, and
is the perturbed matrix, where is a rank-one perturbation.

What is the largest eigenvalue? What about the corresponding eigenvector? Well, the Wigner semicircle law tells us that the noise eigenvalues follow a semicircular distribution, and the signal eigenvalue falls outside the semicircle. If you look at the top eigenvalue as a function of , there’s a critical value of 1 where the top eigenvalue begins to escape the semicircle. Below that critical value, there’s no way to distinguish the signal eigenvalue.

Peche and Feral developed a closed-form formula for the top eigenvalue:
if is greater than one, and
otherwise.

But you can extend this to a broader range of situations. We can consider n x m matrices, and let
and

This models a sample-covariance matrix. Again we have a phase transition:
if is greater than and

Nadakuti’s theorem is in a more general case. If we assume we have a distribution of noise eigenvalues that converges, as the size of the matrix becomes large, to some continuous distribution limited to the interval [a, b] then
if is less than and b otherwise.

Here, refers to the Cauchy transform

The sketch of the proof is fairly simple.
is an eigenvalue of iff 1 is an eigenvalue of

Equivalently,

Equivalently, the eigenvalues satisfy
. Since the are assumed to be distributed uniformly on the sphere, this means that

So the moral of the story is that we can determine, rather specifically, at what point a set of data is too noisy for dimensionality reduction techniques to have any hope of accuracy. There are analogues to this in the work on sparse approximation of Donoho and Tanner, and in free probability theory.

Edit: Here’s the link to the paper.

## Random Matrix Theory in New ScientistApril 14, 2010

Posted by Sarah in Uncategorized.
Tags: