Searcharxiv⌕ Search

arXiv subjects

Claire Mathieu

Publications and source records attributed to Claire Mathieu.

At least 37 records · Page 2Linked to original sources
arXiv2018

Instance-Optimality in the Noisy Value-and Comparison-Model --- Accept, Accept, Strong Accept: Which Papers get in?

Motivated by crowdsourced computation, peer-grading, and recommendation systems, Braverman, Mao and Weinberg [STOC'16] studied the \emph{query} and \emph{round} complexity of fundamental problems such as finding the maximum (\textsc{max}), finding all elements above a certain value (\textsc{threshold-$v$}) or computing the top$-k$ elements (\textsc{Top}-$k$) in a noisy environment. For example, consider the task of selecting papers for a conference. This task is challenging due the crowdsourcing nature of peer reviews: the results of reviews are noisy and it is necessary to parallelize the review process as much as possible. We study the noisy value model and the noisy comparison model: In the \emph{noisy value model}, a reviewer is asked to evaluate a single element: "What is the value of paper $i$?" (\eg accept). In the \emph{noisy comparison model} (introduced in the seminal work of Feige, Peleg, Raghavan and Upfal [SICOMP'94]) a reviewer is asked to do a pairwise comparison: "Is paper $i$ better than paper $j$?" In this paper, we show optimal worst-case query complexity for the \textsc{max},\textsc{threshold-$v$} and \textsc{Top}-$k$ problems. For \textsc{max} and \textsc{Top}-$k$, we obtain optimal worst-case upper and lower bounds on the round vs query complexity in both models. For \textsc{threshold}-$v$, we obtain optimal query complexity and nearly-optimal round complexity, where $k$ is the size of the output) for both models. We then go beyond the worst-case and address the question of the importance of knowledge of the instance by providing, for a large range of parameters, instance-optimal algorithms with respect to the query complexity. Furthermore, we show that the value model is strictly easier than the comparison model.

cs.DS↗
arXiv2018

Polarization dependent chemistry of ferroelectric BaTiO3 (001) domains

Recent works suggest that the surface chemistry, in particular, the presence of oxygen vacancies can affect the polarization in a ferroelectric material. This should, in turn, influence the domain ordering driven by the need to screen the depolarizing field. Here we show using density functional theory that the presence of oxygen vacancies at the surface of BaTiO3 (001) preferentially stabilizes an inward pointing, P-, polarization. Mirror electron microscopy measurements of the domain ordering confirm the theoretical results.

cond-mat.mtrl-sci↗
arXiv2017

Dynamic clustering to minimize the sum of radii

In this paper, we study the problem of opening centers to cluster a set of clients in a metric space so as to minimize the sum of the costs of the centers and of the cluster radii, in a dynamic environment where clients arrive and depart, and the solution must be updated efficiently while remaining competitive with respect to the current optimal solution. We call this dynamic sum-of-radii clustering problem. We present a data structure that maintains a solution whose cost is within a constant factor of the cost of an optimal solution in metric spaces with bounded doubling dimension and whose worst-case update time is logarithmic in the parameters of the problem.

cs.DS↗
arXiv2017

Hierarchical Clustering: Objective Functions and Algorithms

Hierarchical clustering is a recursive partitioning of a dataset into clusters at an increasingly finer granularity. Motivated by the fact that most work on hierarchical clustering was based on providing algorithms, rather than optimizing a specific objective, Dasgupta framed similarity-based hierarchical clustering as a combinatorial optimization problem, where a `good' hierarchical clustering is one that minimizes some cost function. He showed that this cost function has certain desirable properties. We take an axiomatic approach to defining `good' objective functions for both similarity and dissimilarity-based hierarchical clustering. We characterize a set of "admissible" objective functions (that includes Dasgupta's one) that have the property that when the input admits a `natural' hierarchical clustering, it has an optimal value. Equipped with a suitable objective function, we analyze the performance of practical algorithms, as well as develop better algorithms. For similarity-based hierarchical clustering, Dasgupta showed that the divisive sparsest-cut approach achieves an $O(\log^{3/2} n)$-approximation. We give a refined analysis of the algorithm and show that it in fact achieves an $O(\sqrt{\log n})$-approx. (Charikar and Chatziafratis independently proved that it is a $O(\sqrt{\log n})$-approx.). This improves upon the LP-based $O(\log n)$-approx. of Roy and Pokutta. For dissimilarity-based hierarchical clustering, we show that the classic average-linkage algorithm gives a factor 2 approx., and provide a simple and better algorithm that gives a factor 3/2 approx.. Finally, we consider `beyond-worst-case' scenario through a generalisation of the stochastic block model for hierarchical clustering. We show that Dasgupta's cost function has desirable properties for these inputs and we provide a simple 1 + o(1)-approximation in this setting.

cs.DS↗
arXiv2016

Optimization of Bootstrapping in Circuits

In 2009, Gentry proposed the first Fully Homomorphic Encryption (FHE) scheme, an extremely powerful cryptographic primitive that enables to perform computations, i.e., to evaluate circuits, on encrypted data without decrypting them first. This has many applications, in particular in cloud computing. In all currently known FHE schemes, encryptions are associated to some (non-negative integer) noise level, and at each evaluation of an AND gate, the noise level increases. This is problematic because decryption can only work if the noise level stays below some maximum level $L$ at every gate of the circuit. To ensure that property, it is possible to perform an operation called \emph{bootstrapping} to reduce the noise level. However, bootstrapping is time-consuming and has been identified as a critical operation. This motivates a new problem in discrete optimization, that of choosing where in the circuit to perform bootstrapping operations so as to control the noise level; the goal is to minimize the number of bootstrappings in circuits. In this paper, we formally define the \emph{bootstrap problem}, we design a polynomial-time $L$-approximation algorithm using a novel method of rounding of a linear program, and we show a matching hardness result: $(L-ε)$-inapproximability for any $ε>0$.

cs.DS↗
arXiv2016

Local search yields approximation schemes for k-means and k-median in Euclidean and minor-free metrics

We give the first polynomial-time approximation schemes (PTASs) for the following problems: (1) uniform facility location in edge-weighted planar graphs; (2) $k$-median and $k$-means in edge-weighted planar graphs; (3) $k$-means in Euclidean spaces of bounded dimension. Our first and second results extend to minor-closed families of graphs. All our results extend to cost functions that are the $p$-th power of the shortest-path distance. The algorithm is local search where the local neighborhood of a solution $S$ consists of all solutions obtained from $S$ by removing and adding $1/ε^{O(1)}$ centers.

cs.DS↗
arXiv2015

The Unreasonable Success of Local Search: Geometric Optimization

What is the effectiveness of local search algorithms for geometric problems in the plane? We prove that local search with neighborhoods of magnitude $1/ε^c$ is an approximation scheme for the following problems in the Euclidian plane: TSP with random inputs, Steiner tree with random inputs, facility location (with worst case inputs), and bicriteria $k$-median (also with worst case inputs). The randomness assumption is necessary for TSP.

cs.CG↗
arXiv2015

Bigtable Merge Compaction

NoSQL databases are widely used for massive data storage and real-time web applications. Yet important aspects of these data structures are not well understood. For example, NoSQL databases write most of their data to a collection of files on disk, meanwhile periodically compacting subsets of these files. A compaction policy must choose which files to compact, and when to compact them, without knowing the future workload. Although these choices can affect computational efficiency by orders of magnitude, existing literature lacks tools for designing and analyzing online compaction policies --- policies are now chosen largely by trial and error. Here we introduce tools for the design and analysis of compaction policies for Google Bigtable, propose new policies, give average-case and worst-case competitive analyses, and present preliminary empirical benchmarks.

cs.DS↗
arXiv2015

Near-Linear Query Complexity for Graph Inference

How efficiently can we find an unknown graph using distance or shortest path queries between its vertices? Let $G = (V,E)$ be an unweighted, connected graph of bounded degree. The edge set $E$ is initially unknown, and the graph can be accessed using a \emph{distance oracle}, which receives a pair of vertices $(u,v)$ and returns the distance between $u$ and $v$. In the \emph{verification} problem, we are given a hypothetical graph $\hat G = (V,\hat E)$ and want to check whether $G$ is equal to $\hat G$. We analyze a natural greedy algorithm and prove that it uses $n^{1+o(1)}$ distance queries. In the more difficult \emph{reconstruction} problem, $\hat G$ is not given, and the goal is to find the graph $G$. If the graph can be accessed using a \emph{shortest path oracle}, which returns not just the distance but an actual shortest path between $u$ and $v$, we show that extending the idea of greedy gives a reconstruction algorithm that uses $n^{1+o(1)}$ shortest path queries. When the graph has bounded treewidth, we further bound the query complexity of the greedy algorithms for both problems by $\tilde O(n)$. When the graph is chordal, we provide a randomized algorithm for reconstruction using $\tilde O(n)$ distance queries.

cs.DS↗
arXiv2014

Maximum Matching in Semi-Streaming with Few Passes

In the semi-streaming model, an algorithm receives a stream of edges of a graph in arbitrary order and uses a memory of size $O(n \mbox{ polylog } n)$, where $n$ is the number of vertices of a graph. In this work, we present semi-streaming algorithms that perform one or two passes over the input stream for maximum matching with no restrictions on the input graph, and for the important special case of bipartite graphs that we refer to as maximum bipartite matching (MBM). The Greedy matching algorithm performs one pass over the input and outputs a $1/2$ approximation. Whether there is a better one-pass algorithm has been an open question since the appearance of the first paper on streaming algorithms for matching problems in 2005 [Feigenbaum et al., SODA 2005]. We make the following progress on this problem: In the one-pass setting, we show that there is a deterministic semi-streaming algorithm for MBM with expected approximation factor $1/2+0.005$, assuming that edges arrive one by one in (uniform) random order. We extend this algorithm to general graphs, and we obtain a $1/2+0.003$ approximation. In the two-pass setting, we do not require the random arrival order assumption (the edge stream is in arbitrary order). We present a simple randomized two-pass semi-streaming algorithm for MBM with expected approximation factor $1/2 + 0.019$. Furthermore, we discuss a more involved deterministic two-pass semi-streaming algorithm for MBM with approximation factor $1/2 + 0.019$ and a generalization of this algorithm to general graphs with approximation factor $1/2 + 0.0071$.

cs.DS↗
arXiv2014

Facility Location in Evolving Metrics

Understanding the dynamics of evolving social or infrastructure networks is a challenge in applied areas such as epidemiology, viral marketing, or urban planning. During the past decade, data has been collected on such networks but has yet to be fully analyzed. We propose to use information on the dynamics of the data to find stable partitions of the network into groups. For that purpose, we introduce a time-dependent, dynamic version of the facility location problem, that includes a switching cost when a client's assignment changes from one facility to another. This might provide a better representation of an evolving network, emphasizing the abrupt change of relationships between subjects rather than the continuous evolution of the underlying network. We show that in realistic examples this model yields indeed better fitting solutions than optimizing every snapshot independently. We present an $O(\log nT)$-approximation algorithm and a matching hardness result, where $n$ is the number of clients and $T$ the number of time steps. We also give an other algorithms with approximation ratio $O(\log nT)$ for the variant where one pays at each time step (leasing) for each open facility.

cs.SI↗
arXiv2014

A polynomial-time approximation scheme for Euclidean Steiner forest

We give a randomized O(n polylog n)-time approximation scheme for the Steiner forest problem in the Euclidean plane. For every fixed eps > 0 and given n terminals in the plane with connection requests between some pairs of terminals, our scheme finds a (1 + eps)-approximation to the minimum-length forest that connects every requested pair of terminals.

cs.CG↗
arXiv2014

Energy-efficient algorithms for non-preemptive speed-scaling

We improve complexity bounds for energy-efficient speed scheduling problems for both the single processor and multi-processor cases. Energy conservation has become a major concern, so revisiting traditional scheduling problems to take into account the energy consumption has been part of the agenda of the scheduling community for the past few years. We consider the energy minimizing speed scaling problem introduced by Yao et al. where we wish to schedule a set of jobs, each with a release date, deadline and work volume, on a set of identical processors. The processors may change speed as a function of time and the energy they consume is the $α$th power of its speed. The objective is then to find a feasible schedule which minimizes the total energy used. We show that in the setting with an arbitrary number of processors where all work volumes are equal, there is a $2(1+\varepsilon)(5(1+\varepsilon))^{α-1}\tilde{B}_α=O_α(1)$ approximation algorithm, where $\tilde{B}_α$ is the generalized Bell number. This is the first constant factor algorithm for this problem. This algorithm extends to general unequal processor-dependent work volumes, up to losing a factor of $(\frac{(1+r)r}{2})^α$ in the approximation, where $r$ is the maximum ratio between two work volumes. We then show this latter problem is APX-hard, even in the special case when all release dates and deadlines are equal and $r$ is 4. In the single processor case, we introduce a new linear programming formulation of speed scaling and prove that its integrality gap is at most $12^{α-1}$. As a corollary, we obtain a $(12(1+\varepsilon))^{α-1}$ approximation algorithm where there is a single processor, improving on the previous best bound of $2^{α-1}(1+\varepsilon)^α\tilde{B}_α$ when $α\ge 25$.

cs.DS↗
arXiv2013

First-Come-First-Served for Online Slot Allocation and Huffman Coding

Can one choose a good Huffman code on the fly, without knowing the underlying distribution? Online Slot Allocation (OSA) models this and similar problems: There are n slots, each with a known cost. There are n items. Requests for items are drawn i.i.d. from a fixed but hidden probability distribution p. After each request, if the item, i, was not previously requested, then the algorithm (knowing the slot costs and the requests so far, but not p) must place the item in some vacant slot j(i). The goal is to minimize the sum, over the items, of the probability of the item times the cost of its assigned slot. The optimal offline algorithm is trivial: put the most probable item in the cheapest slot, the second most probable item in the second cheapest slot, etc. The optimal online algorithm is First Come First Served (FCFS): put the first requested item in the cheapest slot, the second (distinct) requested item in the second cheapest slot, etc. The optimal competitive ratios for any online algorithm are 1+H(n-1) ~ ln n for general costs and 2 for concave costs. For logarithmic costs, the ratio is, asymptotically, 1: FCFS gives cost opt + O(log opt). For Huffman coding, FCFS yields an online algorithm (one that allocates codewords on demand, without knowing the underlying probability distribution) that guarantees asymptotically optimal cost: at most opt + 2 log(1+opt) + 2.

cs.DS↗
arXiv2013

The min mean-weight cycle in a random network

The mean weight of a cycle in an edge-weighted graph is the sum of the cycle's edge weights divided by the cycle's length. We study the minimum mean-weight cycle on the complete graph on n vertices, with random i.i.d. edge weights drawn from an exponential distribution with mean 1. We show that the probability of the min mean weight being at most c/n tends to a limiting function of c which is analytic for c<=1/e, discontinuous at c=1/e, and equal to 1 for c>1/e. We further show that if the min mean weight is <=1/(en), then the length of the relevant cycle is Theta_p(1) (i.e., it has a limiting probability distribution which does not scale with n), but that if the min mean weight is >1/(en), then the relevant cycle almost always has mean weight (1+o(1))/(en) and length at least (2/pi^2-o(1)) log^2 n log log n.

math.PR↗
arXiv2013

Graph Reconstruction via Distance Oracles

We study the problem of reconstructing a hidden graph given access to a distance oracle. We design randomized algorithms for the following problems: reconstruction of a degree bounded graph with query complexity $\tilde{O}(n^{3/2})$; reconstruction of a degree bounded outerplanar graph with query complexity $\tilde{O}(n)$; and near-optimal approximate reconstruction of a general graph.

cs.DS↗
arXiv2012

Huffman Coding with Letter Costs: A Linear-Time Approximation Scheme

We give a polynomial-time approximation scheme for the generalization of Huffman Coding in which codeword letters have non-uniform costs (as in Morse code, where the dash is twice as long as the dot). The algorithm computes a (1+epsilon)-approximate solution in time O(n + f(epsilon) log^3 n), where n is the input size.

cs.DS↗