arXiv subjects

Petros Drineas

Publications and source records attributed to Petros Drineas.

At least 55 records · Page 3Linked to original sources

arXiv2015

Feature Selection for Linear SVM with Provable Guarantees

We give two provably accurate feature-selection techniques for the linear SVM. The algorithms run in deterministic and randomized time respectively. Our algorithms can be used in an unsupervised or supervised setting. The supervised approach is based on sampling features from support vectors. We prove that the margin in the feature space is preserved to within $ε$-relative error of the margin in the full feature space in the worst-case. In the unsupervised setting, we also provide worst-case guarantees of the radius of the minimum enclosing ball, thereby ensuring comparable generalization as in the full feature space and resolving an open problem posed in Dasgupta et al. We present extensive experiments on real-world datasets to support our theory and to demonstrate that our method is competitive and often better than prior state-of-the-art, for which there are no known provable guarantees.

stat.ML↗

arXiv2015

Tensor sparsification via a bound on the spectral norm of random tensors

Given an order-$d$ tensor $\tensor A \in \R^{n \times n \times...\times n}$, we present a simple, element-wise sparsification algorithm that zeroes out all sufficiently small elements of $\tensor A$, keeps all sufficiently large elements of $\tensor A$, and retains some of the remaining elements with probabilities proportional to the square of their magnitudes. We analyze the approximation accuracy of the proposed algorithm using a powerful inequality that we derive. This inequality bounds the spectral norm of a random tensor and is of independent interest. As a result, we obtain novel bounds for the tensor sparsification problem.

math.NA↗

arXiv2014

Randomized Dimensionality Reduction for k-means Clustering

We study the topic of dimensionality reduction for $k$-means clustering. Dimensionality reduction encompasses the union of two approaches: \emph{feature selection} and \emph{feature extraction}. A feature selection based algorithm for $k$-means clustering selects a small subset of the input features and then applies $k$-means clustering on the selected features. A feature extraction based algorithm for $k$-means clustering constructs a small set of new artificial features and then applies $k$-means clustering on the constructed features. Despite the significance of $k$-means clustering as well as the wealth of heuristic methods addressing it, provably accurate feature selection methods for $k$-means clustering are not known. On the other hand, two provably accurate feature extraction methods for $k$-means clustering are known in the literature; one is based on random projections and the other is based on the singular value decomposition (SVD). This paper makes further progress towards a better understanding of dimensionality reduction for $k$-means clustering. Namely, we present the first provably accurate feature selection method for $k$-means clustering and, in addition, we present two feature extraction methods. The first feature extraction method is based on random projections and it improves upon the existing results in terms of time complexity and number of features needed to be extracted. The second feature extraction method is based on fast approximate SVD factorizations and it also improves upon the existing results in terms of time complexity. The proposed algorithms are randomized and provide constant-factor approximation guarantees with respect to the optimal $k$-means objective value.

cs.DS↗

arXiv2014

Random Projections for Linear Support Vector Machines

Let X be a data matrix of rank ρ, whose rows represent n points in d-dimensional space. The linear support vector machine constructs a hyperplane separator that maximizes the 1-norm soft margin. We develop a new oblivious dimension reduction technique which is precomputed and can be applied to any input matrix X. We prove that, with high probability, the margin and minimum enclosing ball in the feature space are preserved to within ε-relative error, ensuring comparable generalization as in the original space in the case of classification. For regression, we show that the margin is preserved to ε-relative error with high probability. We present extensive experiments with real and synthetic data to support our theory.

cs.LG↗

arXiv2014

The Fast Cauchy Transform and Faster Robust Linear Regression

We provide fast algorithms for overconstrained $\ell_p$ regression and related problems: for an $n\times d$ input matrix $A$ and vector $b\in\mathbb{R}^n$, in $O(nd\log n)$ time we reduce the problem $\min_{x\in\mathbb{R}^d} \|Ax-b\|_p$ to the same problem with input matrix $\tilde A$ of dimension $s \times d$ and corresponding $\tilde b$ of dimension $s\times 1$. Here, $\tilde A$ and $\tilde b$ are a coreset for the problem, consisting of sampled and rescaled rows of $A$ and $b$; and $s$ is independent of $n$ and polynomial in $d$. Our results improve on the best previous algorithms when $n\gg d$, for all $p\in[1,\infty)$ except $p=2$. We also provide a suite of improved results for finding well-conditioned bases via ellipsoidal rounding, illustrating tradeoffs between running time and conditioning quality, including a one-pass conditioning algorithm for general $\ell_p$ problems. We also provide an empirical evaluation of implementations of our algorithms for $p=1$, comparing them with related algorithms. Our empirical results show that, in the asymptotic regime, the theory is a very good guide to the practical performance of these algorithms. Our algorithms use our faster constructions of well-conditioned bases for $\ell_p$ spaces and, for $p=1$, a fast subspace embedding of independent interest that we call the Fast Cauchy Transform: a distribution over matrices $Π:\mathbb{R}^n\mapsto \mathbb{R}^{O(d\log d)}$, found obliviously to $A$, that approximately preserves the $\ell_1$ norms: that is, with large probability, simultaneously for all $x$, $\|Ax\|_1 \approx \|ΠAx\|_1$, with distortion $O(d^{2+η})$, for an arbitrarily small constant $η>0$; and, moreover, $ΠA$ can be computed in $O(nd\log d)$ time. The techniques underlying our Fast Cauchy Transform include fast Johnson-Lindenstrauss transforms, low-coherence matrices, and rescaling by Cauchy random variables.

cs.DS↗

arXiv2014

A Note on Randomized Element-wise Matrix Sparsification

Given a matrix A \in R^{m x n}, we present a randomized algorithm that sparsifies A by retaining some of its elements by sampling them according to a distribution that depends on both the square and the absolute value of the entries. We combine the ideas of [4, 1] and provide an elementary proof of the approximation accuracy of our algorithm following [4] without the truncation step.

cs.IT↗

arXiv2013

Identifying Influential Entries in a Matrix

For any matrix A in R^(m x n) of rank ρ, we present a probability distribution over the entries of A (the element-wise leverage scores of equation (2)) that reveals the most influential entries in the matrix. From a theoretical perspective, we prove that sampling at most s = O ((m + n) ρ^2 ln (m + n)) entries of the matrix (see eqn. (3) for the precise value of s) with respect to these scores and solving the nuclear norm minimization problem on the sampled entries, reconstructs A exactly. To the best of our knowledge, these are the strongest theoretical guarantees on matrix completion without any incoherence assumptions on the matrix A. From an experimental perspective, we show that entries corresponding to high element-wise leverage scores reveal structural properties of the data matrix that are of interest to domain scientists.

math.NA↗

arXiv2013

Near-Optimal Column-Based Matrix Reconstruction

We consider low-rank reconstruction of a matrix using its columns and we present asymptotically optimal algorithms for both spectral norm and Frobenius norm reconstruction. The main tools we introduce to obtain our r esults are: (i) the use of fast approximate SVD-like decompositions for column reconstruction, and (ii) two deter ministic algorithms for selecting rows from matrices with orthonormal columns, building upon the sparse represen tation theorem for decompositions of the identity that appeared in \cite{BSS09}.

cs.DS↗

arXiv2013

Near-optimal Coresets For Least-Squares Regression

We study (constrained) least-squares regression as well as multiple response least-squares regression and ask the question of whether a subset of the data, a coreset, suffices to compute a good approximate solution to the regression. We give deterministic, low order polynomial-time algorithms to construct such coresets with approximation guarantees, together with lower bounds indicating that there is not much room for improvement upon our results.

cs.DS↗

arXiv2012

Fast approximation of matrix coherence and statistical leverage

The statistical leverage scores of a matrix $A$ are the squared row-norms of the matrix containing its (top) left singular vectors and the coherence is the largest leverage score. These quantities are of interest in recently-popular problems such as matrix completion and Nyström-based low-rank matrix approximation as well as in large-scale statistical data analysis applications more generally; moreover, they are of interest since they define the key structural nonuniformity that must be dealt with in developing fast randomized matrix algorithms. Our main result is a randomized algorithm that takes as input an arbitrary $n \times d$ matrix $A$, with $n \gg d$, and that returns as output relative-error approximations to all $n$ of the statistical leverage scores. The proposed algorithm runs (under assumptions on the precise values of $n$ and $d$) in $O(n d \log n)$ time, as opposed to the $O(nd^2)$ time required by the naïve algorithm that involves computing an orthogonal basis for the range of $A$. Our analysis may be viewed in terms of computing a relative-error approximation to an underconstrained least-squares approximation problem, or, relatedly, it may be viewed as an application of Johnson-Lindenstrauss type ideas. Several practically-important extensions of our basic result are also described, including the approximation of so-called cross-leverage scores, the extension of these ideas to matrices with $n \approx d$, and the extension to streaming environments.

cs.DS↗

arXiv2011

A Note on Element-wise Matrix Sparsification via a Matrix-valued Bernstein Inequality

Given an n x n matrix A, we present a simple, element-wise sparsification algorithm that zeroes out all sufficiently small elements of A and then retains some of the remaining elements with probabilities proportional to the square of their magnitudes. We analyze the approximation accuracy of the proposed algorithm using a recent, elegant non-commutative Bernstein inequality, and compare our bounds with all existing (to the best of our knowledge) element-wise matrix sparsification algorithms.

cs.DS↗

arXiv2010

Random Projections for $k$-means Clustering

This paper discusses the topic of dimensionality reduction for $k$-means clustering. We prove that any set of $n$ points in $d$ dimensions (rows in a matrix $A \in \RR^{n \times d}$) can be projected into $t = Ω(k / \eps^2)$ dimensions, for any $\eps \in (0,1/3)$, in $O(n d \lceil \eps^{-2} k/ \log(d) \rceil )$ time, such that with constant probability the optimal $k$-partition of the point set is preserved within a factor of $2+\eps$. The projection is done by post-multiplying $A$ with a $d \times t$ random matrix $R$ having entries $+1/\sqrt{t}$ or $-1/\sqrt{t}$ with equal probability. A numerical implementation of our technique and experiments on a large face images dataset verify the speed and the accuracy of our theoretical results.

cs.AI↗

arXiv2010

Faster Least Squares Approximation

Least squares approximation is a technique to find an approximate solution to a system of linear equations that has no exact solution. In a typical setting, one lets $n$ be the number of constraints and $d$ be the number of variables, with $n \gg d$. Then, existing exact methods find a solution vector in $O(nd^2)$ time. We present two randomized algorithms that provide very accurate relative-error approximations to the optimal value and the solution vector of a least squares approximation problem more rapidly than existing exact algorithms. Both of our algorithms preprocess the data with the Randomized Hadamard Transform. One then uniformly randomly samples constraints and solves the smaller problem on those constraints, and the other performs a sparse random projection and solves the smaller problem on those projected coordinates. In both cases, solving the smaller problem provides relative-error approximations, and, if $n$ is sufficiently larger than $d$, the approximate solution can be computed in $O(nd \log d)$ time.

cs.DS↗

arXiv2010

Effective Resistances, Statistical Leverage, and Applications to Linear Equation Solving

Recent work in theoretical computer science and scientific computing has focused on nearly-linear-time algorithms for solving systems of linear equations. While introducing several novel theoretical perspectives, this work has yet to lead to practical algorithms. In an effort to bridge this gap, we describe in this paper two related results. Our first and main result is a simple algorithm to approximate the solution to a set of linear equations defined by a Laplacian (for a graph $G$ with $n$ nodes and $m \le n^2$ edges) constraint matrix. The algorithm is a non-recursive algorithm; even though it runs in $O(n^2 \cdot \polylog(n))$ time rather than $O(m \cdot polylog(n))$ time (given an oracle for the so-called statistical leverage scores), it is extremely simple; and it can be used to compute an approximate solution with a direct solver. In light of this result, our second result is a straightforward connection between the concept of graph resistance (which has proven useful in recent algorithms for linear equation solvers) and the concept of statistical leverage (which has proven useful in numerically-implementable randomized algorithms for large matrix problems and which has a natural data-analytic interpretation).

math.NA↗

arXiv2010

An Improved Approximation Algorithm for the Column Subset Selection Problem

We consider the problem of selecting the best subset of exactly $k$ columns from an $m \times n$ matrix $A$. We present and analyze a novel two-stage algorithm that runs in $O(\min\{mn^2,m^2n\})$ time and returns as output an $m \times k$ matrix $C$ consisting of exactly $k$ columns of $A$. In the first (randomized) stage, the algorithm randomly selects $Θ(k \log k)$ columns according to a judiciously-chosen probability distribution that depends on information in the top-$k$ right singular subspace of $A$. In the second (deterministic) stage, the algorithm applies a deterministic column-selection procedure to select and return exactly $k$ columns from the set of columns selected in the first stage. Let $C$ be the $m \times k$ matrix containing those $k$ columns, let $P_C$ denote the projection matrix onto the span of those columns, and let $A_k$ denote the best rank-$k$ approximation to the matrix $A$. Then, we prove that, with probability at least 0.8, $$ \FNorm{A - P_CA} \leq Θ(k \log^{1/2} k) \FNorm{A-A_k}. $$ This Frobenius norm bound is only a factor of $\sqrt{k \log k}$ worse than the best previously existing existential result and is roughly $O(\sqrt{k!})$ better than the best previous algorithmic result for the Frobenius norm version of this Column Subset Selection Problem (CSSP). We also prove that, with probability at least 0.8, $$ \TNorm{A - P_CA} \leq Θ(k \log^{1/2} k)\TNorm{A-A_k} + Θ(k^{3/4}\log^{1/4}k)\FNorm{A-A_k}. $$ This spectral norm bound is not directly comparable to the best previously existing bounds for the spectral norm version of this CSSP. Our bound depends on $\FNorm{A-A_k}$, whereas previous results depend on $\sqrt{n-k}\TNorm{A-A_k}$; if these two quantities are comparable, then our bound is asymptotically worse by a $(k \log k)^{1/4}$ factor.

cs.DS↗

arXiv2009

Random Projections for the Nonnegative Least-Squares Problem

Constrained least-squares regression problems, such as the Nonnegative Least Squares (NNLS) problem, where the variables are restricted to take only nonnegative values, often arise in applications. Motivated by the recent development of the fast Johnson-Lindestrauss transform, we present a fast random projection type approximation algorithm for the NNLS problem. Our algorithm employs a randomized Hadamard transform to construct a much smaller NNLS problem and solves this smaller problem using a standard NNLS solver. We prove that our approach finds a nonnegative solution vector that, with high probability, is close to the optimum nonnegative solution in a relative error approximation sense. We experimentally evaluate our approach on a large collection of term-document data and verify that it does offer considerable speedups without a significant loss in accuracy. Our analysis is based on a novel random projection type result that might be of independent interest. In particular, given a tall and thin matrix $Φ\in \mathbb{R}^{n \times d}$ ($n \gg d$) and a vector $y \in \mathbb{R}^d$, we prove that the Euclidean length of $Φy$ can be estimated very accurately by the Euclidean length of $\tildeΦy$, where $\tildeΦ$ consists of a small subset of (appropriately rescaled) rows of $Φ$.

cs.DS↗

arXiv2007

Relative-Error CUR Matrix Decompositions

Many data analysis applications deal with large matrices and involve approximating the matrix using a small number of ``components.'' Typically, these components are linear combinations of the rows and columns of the matrix, and are thus difficult to interpret in terms of the original features of the input data. In this paper, we propose and study matrix approximations that are explicitly expressed in terms of a small number of columns and/or rows of the data matrix, and thereby more amenable to interpretation in terms of the original data. Our main algorithmic results are two randomized algorithms which take as input an $m \times n$ matrix $A$ and a rank parameter $k$. In our first algorithm, $C$ is chosen, and we let $A'=CC^+A$, where $C^+$ is the Moore-Penrose generalized inverse of $C$. In our second algorithm $C$, $U$, $R$ are chosen, and we let $A'=CUR$. ($C$ and $R$ are matrices that consist of actual columns and rows, respectively, of $A$, and $U$ is a generalized inverse of their intersection.) For each algorithm, we show that with probability at least $1-δ$: $$ ||A-A'||_F \leq (1+ε) ||A-A_k||_F, $$ where $A_k$ is the ``best'' rank-$k$ approximation provided by truncating the singular value decomposition (SVD) of $A$. The number of columns of $C$ and rows of $R$ is a low-degree polynomial in $k$, $1/ε$, and $\log(1/δ)$. Our two algorithms are the first polynomial time algorithms for such low-rank matrix approximations that come with relative-error guarantees; previously, in some cases, it was not even known whether such matrix decompositions exist. Both of our algorithms are simple, they take time of the order needed to approximately compute the top $k$ singular vectors of $A$, and they use a novel, intuitive sampling method called ``subspace sampling.''

cs.DS↗

arXiv2007

Sampling Algorithms and Coresets for Lp Regression

The Lp regression problem takes as input a matrix $A \in \Real^{n \times d}$, a vector $b \in \Real^n$, and a number $p \in [1,\infty)$, and it returns as output a number ${\cal Z}$ and a vector $x_{opt} \in \Real^d$ such that ${\cal Z} = \min_{x \in \Real^d} ||Ax -b||_p = ||Ax_{opt}-b||_p$. In this paper, we construct coresets and obtain an efficient two-stage sampling-based approximation algorithm for the very overconstrained ($n \gg d$) version of this classical problem, for all $p \in [1, \infty)$. The first stage of our algorithm non-uniformly samples $\hat{r}_1 = O(36^p d^{\max\{p/2+1, p\}+1})$ rows of $A$ and the corresponding elements of $b$, and then it solves the Lp regression problem on the sample; we prove this is an 8-approximation. The second stage of our algorithm uses the output of the first stage to resample $\hat{r}_1/ε^2$ constraints, and then it solves the Lp regression problem on the new sample; we prove this is a $(1+ε)$-approximation. Our algorithm unifies, improves upon, and extends the existing algorithms for special cases of Lp regression, namely $p = 1,2$. In course of proving our result, we develop two concepts--well-conditioned bases and subspace-preserving sampling--that are of independent interest.

cs.DS↗