arXiv subjects

Anastasios Zouzias

Publications and source records attributed to Anastasios Zouzias.

23 records · Page 2Linked to original sources

arXiv2011

On the Certification of the Restricted Isometry Property

Compressed sensing is a technique for finding sparse solutions to underdetermined linear systems. This technique relies on properties of the sensing matrix such as the restricted isometry property. Sensing matrices that satisfy the restricted isometry property with optimal parameters are mainly obtained via probabilistic arguments. Given any matrix, deciding whether it satisfies the restricted isometry property is a non-trivial computational problem. In this paper, we give reductions from dense subgraph problems to the certification of the restricted isometry property. This gives evidence that certifying the restricted isometry property is unlikely to be feasible in polynomial-time. Moreover, on the positive side we propose an improvement on the brute-force enumeration algorithm for checking the restricted isometry property. Another contribution of independent interest is a spectral algorithm for certifying that a random graph does not contain any dense k-subgraph. This "skewed spectral algorithm" performs better than the basic spectral algorithm in a certain range of parameters.

cs.CC↗

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

Low Rank Matrix-Valued Chernoff Bounds and Approximate Matrix Multiplication

In this paper we develop algorithms for approximating matrix multiplication with respect to the spectral norm. Let A\in{\RR^{n\times m}} and B\in\RR^{n \times p} be two matrices and \eps>0. We approximate the product A^\top B using two down-sampled sketches, \tilde{A}\in\RR^{t\times m} and \tilde{B}\in\RR^{t\times p}, where t\ll n such that \norm{\tilde{A}^\top \tilde{B} - A^\top B} \leq \eps \norm{A}\norm{B} with high probability. We use two different sampling procedures for constructing \tilde{A} and \tilde{B}; one of them is done by i.i.d. non-uniform sampling rows from A and B and the other is done by taking random linear combinations of their rows. We prove bounds that depend only on the intrinsic dimensionality of A and B, that is their rank and their stable rank; namely the squared ratio between their Frobenius and operator norm. For achieving bounds that depend on rank we employ standard tools from high-dimensional geometry such as concentration of measure arguments combined with elaborate \eps-net constructions. For bounds that depend on the smaller parameter of stable rank this technology itself seems weak. However, we show that in combination with a simple truncation argument is amenable to provide such bounds. To handle similar bounds for row sampling, we develop a novel matrix-valued Chernoff bound inequality which we call low rank matrix-valued Chernoff bound. Thanks to this inequality, we are able to give bounds that depend only on the stable rank of the input matrices...

cs.DS↗

arXiv2010

Low Dimensional Euclidean Volume Preserving Embeddings

Let $\mathcal{P}$ be an $n$-point subset of Euclidean space and $d\geq 3$ be an integer. In this paper we study the following question: What is the smallest (normalized) relative change of the volume of subsets of $\mathcal{P}$ when it is projected into $\RR^d$. We prove that there exists a linear mapping $f:\mathcal{P} \mapsto \RR^d$ that relatively preserves the volume of all subsets of size up to $\lfloor d/2\rfloor$ within at most a factor of $O(n^{2/d}\sqrt{\log n \log\log n})$.

cs.DM↗