arXiv subjects

Michael Saks

Publications and source records attributed to Michael Saks.

28 records · Page 2Linked to original sources

arXiv2013

A Polynomial Time Algorithm for Lossy Population Recovery

We give a polynomial time algorithm for the lossy population recovery problem. In this problem, the goal is to approximately learn an unknown distribution on binary strings of length $n$ from lossy samples: for some parameter $μ$ each coordinate of the sample is preserved with probability $μ$ and otherwise is replaced by a `?'. The running time and number of samples needed for our algorithm is polynomial in $n$ and $1/\varepsilon$ for each fixed $μ>0$. This improves on algorithm of Wigderson and Yehudayoff that runs in quasi-polynomial time for any $μ> 0$ and the polynomial time algorithm of Dvir et al which was shown to work for $μ\gtrapprox 0.30$ by Batman et al. In fact, our algorithm also works in the more general framework of Batman et al. in which there is no a priori bound on the size of the support of the distribution. The algorithm we analyze is implicit in previous work; our main contribution is to analyze the algorithm by showing (via linear programming duality and connections to complex analysis) that a certain matrix associated with the problem has a robust local inverse even though its condition number is exponentially small. A corollary of our result is the first polynomial time algorithm for learning DNFs in the restriction access model of Dvir et al.

cs.DS↗

arXiv2013

Composition limits and separating examples for some Boolean function complexity measures

Block sensitivity ($bs(f)$), certificate complexity ($C(f)$) and fractional certificate complexity ($C^*(f)$) are three fundamental combinatorial measures of complexity of a boolean function $f$. It has long been known that $bs(f) \leq C^{\ast}(f) \leq C(f) =O(bs(f)^2)$. We provide an infinite family of examples for which $C(f)$ grows quadratically in $C^{\ast}(f)$ (and also $bs(f)$) giving optimal separations between these measures. Previously the biggest separation known was $C(f)=C^{\ast}(f)^{\log_{4.5}5}$. We also give a family of examples for which $C^{\ast}(f)=Ω(bs(f)^{3/2})$. These examples are obtained by composing boolean functions in various ways. Here the composition $f \circ g$ of $f$ with $g$ is obtained by substituting for each variable of $f$ a copy of $g$ on disjoint sets of variables. To construct and analyse these examples we systematically investigate the behaviour under function composition of these measures and also the sensitivity measure $s(f)$. The measures $s(f)$, $C(f)$ and $C^{\ast}(f)$ behave nicely under composition: they are submultiplicative (where measure $m$ is submultiplicative if $m(f \circ g) \leq m(f)m(g)$) with equality holding under some fairly general conditions. The measure $bs(f)$ is qualitatively different: it is not submultiplicative. This qualitative difference was not noticed in the previous literature and we correct some errors that appeared in previous papers. We define the composition limit of a measure $m$ at function $f$, $m^{\lim}(f)$ to be the limit as $k$ grows of $m(f^{(k)})^{1/k}$, where $f^{(k)}$ is the iterated composition of $f$ with itself $k$-times. For any function $f$ we show that $bs^{\lim}(f) = (C^*)^{\lim}(f)$ and characterize $s^{\lim}(f), (C^*)^{\lim}(f)$, and $C^{\lim}(f)$ in terms of the largest eigenvalue of a certain set of $2\times 2$ matrices associated with $f$.

cs.CC↗

arXiv2013

Space efficient streaming algorithms for the distance to monotonicity and asymmetric edit distance

Approximating the length of the longest increasing sequence (LIS) of an array is a well-studied problem. We study this problem in the data stream model, where the algorithm is allowed to make a single left-to-right pass through the array and the key resource to be minimized is the amount of additional memory used. We present an algorithm which, for any $δ> 0$, given streaming access to an array of length $n$ provides a $(1+δ)$-multiplicative approximation to the \emph{distance to monotonicity} ($n$ minus the length of the LIS), and uses only $O((\log^2 n)/δ)$ space. The previous best known approximation using polylogarithmic space was a multiplicative 2-factor. Our algorithm can be used to estimate the length of the LIS to within an additive $δn$ for any $δ>0$ while previous algorithms could only achieve additive error $n(1/2-o(1))$. Our algorithm is very simple, being just 3 lines of pseudocode, and has a small update time. It is essentially a polylogarithmic space approximate implementation of a classic dynamic program that computes the LIS. We also give a streaming algorithm for approximating $LCS(x,y)$, the length of the longest common subsequence between strings $x$ and $y$, each of length $n$. Our algorithm works in the asymmetric setting (inspired by \cite{AKO10}), in which we have random access to $y$ and streaming access to $x$, and runs in small space provided that no single symbol appears very often in $y$. More precisely, it gives an additive-$δn$ approximation to $LCS(x,y)$ (and hence also to $E(x,y) = n-LCS(x,y)$, the edit distance between $x$ and $y$ when insertions and deletions, but not substitutions, are allowed), with space complexity $O(k(\log^2 n)/δ)$, where $k$ is the maximum number of times any one symbol appears in $y$.

cs.DS↗

arXiv2012

On Online Labeling with Polynomially Many Labels

In the online labeling problem with parameters n and m we are presented with a sequence of n keys from a totally ordered universe U and must assign each arriving key a label from the label set {1,2,...,m} so that the order of labels (strictly) respects the ordering on U. As new keys arrive it may be necessary to change the labels of some items; such changes may be done at any time at unit cost for each change. The goal is to minimize the total cost. An alternative formulation of this problem is the file maintenance problem, in which the items, instead of being labeled, are maintained in sorted order in an array of length m, and we pay unit cost for moving an item. For the case m=cn for constant c>1, there are known algorithms that use at most O(n log(n)^2) relabelings in total [Itai, Konheim, Rodeh, 1981], and it was shown recently that this is asymptotically optimal [Bulánek, Koucký, Saks, 2012]. For the case of m=Θ(n^C) for C>1, algorithms are known that use O(n log n) relabelings. A matching lower bound was claimed in [Dietz, Seiferas, Zhang, 2004]. That proof involved two distinct steps: a lower bound for a problem they call prefix bucketing and a reduction from prefix bucketing to online labeling. The reduction seems to be incorrect, leaving a (seemingly significant) gap in the proof. In this paper we close the gap by presenting a correct reduction to prefix bucketing. Furthermore we give a simplified and improved analysis of the prefix bucketing lower bound. This improvement allows us to extend the lower bounds for online labeling to the case where the number m of labels is superpolynomial in n. In particular, for superpolynomial m we get an asymptotically optimal lower bound Ω((n log n) / (log log m - log log n)).

cs.DS↗

arXiv2012

Clustering is difficult only when it does not matter

Numerous papers ask how difficult it is to cluster data. We suggest that the more relevant and interesting question is how difficult it is to cluster data sets {\em that can be clustered well}. More generally, despite the ubiquity and the great importance of clustering, we still do not have a satisfactory mathematical theory of clustering. In order to properly understand clustering, it is clearly necessary to develop a solid theoretical basis for the area. For example, from the perspective of computational complexity theory the clustering problem seems very hard. Numerous papers introduce various criteria and numerical measures to quantify the quality of a given clustering. The resulting conclusions are pessimistic, since it is computationally difficult to find an optimal clustering of a given data set, if we go by any of these popular criteria. In contrast, the practitioners' perspective is much more optimistic. Our explanation for this disparity of opinions is that complexity theory concentrates on the worst case, whereas in reality we only care for data sets that can be clustered well. We introduce a theoretical framework of clustering in metric spaces that revolves around a notion of "good clustering". We show that if a good clustering exists, then in many cases it can be efficiently found. Our conclusion is that contrary to popular belief, clustering should not be considered a hard task.

cs.LG↗

arXiv2012

On the practically interesting instances of MAXCUT

The complexity of a computational problem is traditionally quantified based on the hardness of its worst case. This approach has many advantages and has led to a deep and beautiful theory. However, from the practical perspective, this leaves much to be desired. In application areas, practically interesting instances very often occupy just a tiny part of an algorithm's space of instances, and the vast majority of instances are simply irrelevant. Addressing these issues is a major challenge for theoretical computer science which may make theory more relevant to the practice of computer science. Following Bilu and Linial, we apply this perspective to MAXCUT, viewed as a clustering problem. Using a variety of techniques, we investigate practically interesting instances of this problem. Specifically, we show how to solve in polynomial time distinguished, metric, expanding and dense instances of MAXCUT under mild stability assumptions. In particular, $(1+ε)$-stability (which is optimal) suffices for metric and dense MAXCUT. We also show how to solve in polynomial time $Ω(\sqrt{n})$-stable instances of MAXCUT, substantially improving the best previously known result.

cs.CC↗

arXiv2012

A Tail Bound for Read-k Families of Functions

We prove a Chernoff-like large deviation bound on the sum of non-independent random variables that have the following dependence structure. The variables $Y_1,...,Y_r$ are arbitrary Boolean functions of independent random variables $X_1,...,X_m$, modulo a restriction that every $X_i$ influences at most $k$ of the variables $Y_1,...,Y_r$.

cs.DM↗

arXiv2011

Tight lower bounds for online labeling problem

We consider the file maintenance problem (also called the online labeling problem) in which n integer items from the set {1,...,r} are to be stored in an array of size m >= n. The items are presented sequentially in an arbitrary order, and must be stored in the array in sorted order (but not necessarily in consecutive locations in the array). Each new item must be stored in the array before the next item is received. If r<=m then we can simply store item j in location j but if r>m then we may have to shift the location of stored items to make space for a newly arrived item. The algorithm is charged each time an item is stored in the array, or moved to a new location. The goal is to minimize the total number of such moves done by the algorithm. This problem is non-trivial when n= 1, algorithms for this problem with cost O(log(n)^2) per item have been given [IKR81, Wil92, BCD+02]. When m=n, algorithms with cost O(log(n)^3) per item were given [Zha93, BS07]. In this paper we prove lower bounds that show that these algorithms are optimal, up to constant factors. Previously, the only lower bound known for this range of parameters was a lower bound of Ω(log(n)^2) for the restricted class of smooth algorithms [DSZ05a, Zha93]. We also provide an algorithm for the sparse case: If the number of items is polylogarithmic in the array size then the problem can be solved in amortized constant time per item.

cs.DS↗

arXiv2005

Every decision tree has an influential variable

We prove that for any decision tree calculating a boolean function $f:\{-1,1\}^n\to\{-1,1\}$, \[ \Var[f] \le \sum_{i=1}^n δ_i \Inf_i(f), \] where $δ_i$ is the probability that the $i$th input variable is read and $\Inf_i(f)$ is the influence of the $i$th variable on $f$. The variance, influence and probability are taken with respect to an arbitrary product measure on $\{-1,1\}^n$. It follows that the minimum depth of a decision tree calculating a given balanced function is at least the reciprocal of the largest influence of any input variable. Likewise, any balanced boolean function with a decision tree of depth $d$ has a variable with influence at least $\frac{1}{d}$. The only previous nontrivial lower bound known was $Ω(d 2^{-d})$. Our inequality has many generalizations, allowing us to prove influence lower bounds for randomized decision trees, decision trees on arbitrary product probability spaces, and decision trees with non-boolean outputs. As an application of our results we give a very easy proof that the randomized query complexity of nontrivial monotone graph properties is at least $Ω(v^{4/3}/p^{1/3})$, where $v$ is the number of vertices and $p \leq \half$ is the critical threshold probability. This supersedes the milestone $Ω(v^{4/3})$ bound of Hajnal and is sometimes superior to the best known lower bounds of Chakrabarti-Khot and Friedgut-Kahn-Wigderson.

cs.CC↗

arXiv2002

A lower bound on the quantum query complexity of read-once functions

We establish a lower bound of $Ω{(\sqrt{n})}$ on the bounded-error quantum query complexity of read-once Boolean functions, providing evidence for the conjecture that $Ω(\sqrt{D(f)})$ is a lower bound for all Boolean functions. Our technique extends a result of Ambainis, based on the idea that successful computation of a function requires ``decoherence'' of initially coherently superposed inputs in the query register, having different values of the function. The number of queries is bounded by comparing the required total amount of decoherence of a judiciously selected set of input-output pairs to an upper bound on the amount achievable in a single query step. We use an extension of this result to general weights on input pairs, and general superpositions of inputs.

quant-ph↗