arXiv subjects

Lee-Ad Gottlieb

Publications and source records attributed to Lee-Ad Gottlieb.

28 records · Page 2Linked to original sources

arXiv2015

The Traveling Salesman Problem: Low-Dimensionality Implies a Polynomial Time Approximation Scheme

The Traveling Salesman Problem (TSP) is among the most famous NP-hard optimization problems. We design for this problem a randomized polynomial-time algorithm that computes a (1+eps)-approximation to the optimal tour, for any fixed eps>0, in TSP instances that form an arbitrary metric space with bounded intrinsic dimension. The celebrated results of Arora (A-98) and Mitchell (M-99) prove that the above result holds in the special case of TSP in a fixed-dimensional Euclidean space. Thus, our algorithm demonstrates that the algorithmic tractability of metric TSP depends on the dimensionality of the space and not on its specific geometry. This result resolves a problem that has been open since the quasi-polynomial time algorithm of Talwar (T-04).

cs.CC↗

arXiv2015

Adaptive Metric Dimensionality Reduction

We study adaptive data-dependent dimensionality reduction in the context of supervised learning in general metric spaces. Our main statistical contribution is a generalization bound for Lipschitz functions in metric spaces that are doubling, or nearly doubling. On the algorithmic front, we describe an analogue of PCA for metric spaces: namely an efficient procedure that approximates the data's intrinsic dimension, which is often much lower than the ambient dimension. Our approach thus leverages the dual benefits of low dimensionality: (1) more efficient algorithms, e.g., for proximity search, and (2) more optimistic generalization bounds.

cs.LG↗

arXiv2015

Nearly optimal classification for semimetrics

We initiate the rigorous study of classification in semimetric spaces, which are point sets with a distance function that is non-negative and symmetric, but need not satisfy the triangle inequality. For metric spaces, the doubling dimension essentially characterizes both the runtime and sample complexity of classification algorithms --- yet we show that this is not the case for semimetrics. Instead, we define the {\em density dimension} and discover that it plays a central role in the statistical and algorithmic feasibility of learning in semimetric spaces. We present nearly optimal sample compression algorithms and use these to obtain generalization guarantees, including fast rates. The latter hold for general sample compression schemes and may be of independent interest.

cs.LG↗

arXiv2014

Efficient Classification for Metric Data

Recent advances in large-margin classification of data residing in general metric spaces (rather than Hilbert spaces) enable classification under various natural metrics, such as string edit and earthmover distance. A general framework developed for this purpose by von Luxburg and Bousquet [JMLR, 2004] left open the questions of computational efficiency and of providing direct bounds on generalization error. We design a new algorithm for classification in general metric spaces, whose runtime and accuracy depend on the doubling dimension of the data points, and can thus achieve superior classification performance in many common scenarios. The algorithmic core of our approach is an approximate (rather than exact) solution to the classical problems of Lipschitz extension and of Nearest Neighbor Search. The algorithm's generalization performance is guaranteed via the fat-shattering dimension of Lipschitz classifiers, and we present experimental evidence of its superiority to some common kernel methods. As a by-product, we offer a new perspective on the nearest neighbor classifier, which yields significantly sharper risk asymptotics than the classic analysis of Cover and Hart [IEEE Trans. Info. Theory, 1967].

cs.LG↗

arXiv2014

Optimizing Budget Allocation in Graphs

In the classical facility location problem we consider a graph $G$ with fixed weights on the edges of $G$. The goal is then to find an optimal positioning for a set of facilities on the graph with respect to some objective function. We introduce a new framework for facility location problems, where the weights on the graph edges are not fixed, but rather should be assigned. The goal is to find a valid assignment for which the resulting weighted graph optimizes the facility location objective function. We present algorithms for finding the optimal {\em budget allocation} for the center point problem and for the median point problem on trees. Our algorithms run in linear time, both for the case where a candidate vertex is given as part of the input, and for the case where finding a vertex that optimizes the solution is part of the problem. We also present a hardness result for the general graph case of the center point problem, followed by an $O(\log^2(n))$ approximation algorithm on graphs - with general metric spaces.

cs.DS↗

arXiv2014

Light spanners for snowflake metrics

A classic result in the study of spanners is the existence of light low-stretch spanners for Euclidean spaces. These spanners ahve arbitrary low stretch, and weight only a constant factor greater than that of the minimum spanning tree of the points (with dependence on the stretch and Euclidean dimention). A central open problem in this field asks whether other spaces admit low weight spanners as well - for example metric space with low intrinsic dimension - yet only a handful of results of this type are known. In this paper, we consider snowflake metric spaces of low intrinsic dimension. The α-snowflake of a metric (X,δ) is the metric (X,$δ^α$) for 0<α<1. By utilizing an approach completely different than those used for Euclidean spaces, we demonstrate that snowflake metrics admit light spanners. Further, we show that the spanner is of diameter O($\log$n), a result not possible for Euclidean spaces. As an immediate corollary to our spanner, we obtain dramatic improvments in algorithms for the traveling salesman problem in this setting, achieving a polynomial-time approximation scheme with near-linear runtime. Along the way, we show that all ${\ell}_p$ spaces admit light spanners, a result of interest in its own right.

cs.CG↗

arXiv2013

On the Impossibility of Dimension Reduction for Doubling Subsets of $\ell_p$, $p>2$

A major open problem in the field of metric embedding is the existence of dimension reduction for $n$-point subsets of Euclidean space, such that both distortion and dimension depend only on the {\em doubling constant} of the pointset, and not on its cardinality. In this paper, we negate this possibility for $\ell_p$ spaces with $p>2$. In particular, we introduce an $n$-point subset of $\ell_p$ with doubling constant O(1), and demonstrate that any embedding of the set into $\ell_p^d$ with distortion $D$ must have $D\geΩ\left(\left(\frac{c\log n}{d}\right)^{\frac{1}{2}-\frac{1}{p}}\right)$.

cs.CG↗

arXiv2011

VC bounds on the cardinality of nearly orthogonal function classes

We bound the number of nearly orthogonal vectors with fixed VC-dimension over $\setpm^n$. Our bounds are of interest in machine learning and empirical process theory and improve previous bounds by Haussler. The bounds are based on a simple projection argument and the generalize to other product spaces. Along the way we derive tight bounds on the sum of binomial coefficients in terms of the entropy function.

math.CO↗

arXiv2010

Fast, precise and dynamic distance queries

We present an approximate distance oracle for a point set S with n points and doubling dimension λ. For every ε>0, the oracle supports (1+ε)-approximate distance queries in (universal) constant time, occupies space [ε^{-O(λ)} + 2^{O(λ log λ)}]n, and can be constructed in [2^{O(λ)} log3 n + ε^{-O(λ)} + 2^{O(λ log λ)}]n expected time. This improves upon the best previously known constructions, presented by Har-Peled and Mendel. Furthermore, the oracle can be made fully dynamic with expected O(1) query time and only 2^{O(λ)} log n + ε^{-O(λ)} + 2^{O(λ log λ)} update time. This is the first fully dynamic (1+ε)-distance oracle.

cs.DS↗

arXiv2010

Matrix sparsification and the sparse null space problem

We revisit the matrix problems sparse null space and matrix sparsification, and show that they are equivalent. We then proceed to seek algorithms for these problems: We prove the hardness of approximation of these problems, and also give a powerful tool to extend algorithms and heuristics for sparse approximation theory to these problems.

cs.CC↗