Searcharxiv⌕ Search

arXiv subjects

Nick Gravin

Publications and source records attributed to Nick Gravin.

At least 37 records · Page 2Linked to original sources
arXiv2017

Testing Symmetric Markov Chains from a Single Trajectory

Classical distribution testing assumes access to i.i.d. samples from the distribution that is being tested. We initiate the study of Markov chain testing, assuming access to a single trajectory of a Markov Chain. In particular, we observe a single trajectory X0,...,Xt,... of an unknown, symmetric, and finite state Markov Chain M. We do not control the starting state X0, and we cannot restart the chain. Given our single trajectory, the goal is to test whether M is identical to a model Markov Chain M0 , or far from it under an appropriate notion of difference. We propose a measure of difference between two Markov chains, motivated by the early work of Kazakos [Kaz78], which captures the scaling behavior of the total variation distance between trajectories sampled from the Markov chains as the length of these trajectories grows. We provide efficient testers and information-theoretic lower bounds for testing identity of symmetric Markov chains under our proposed measure of difference, which are tight up to logarithmic factors if the hitting times of the model chain M0 is O(n) in the size of the state space n.

cs.LG↗
arXiv2017

On moments of a polytope

We show that the multivariate generating function of appropriately normalized moments of a measure with homogeneous polynomial density supported on a compact polytope P in R^d is a rational function. Its denominator is the product of linear forms dual to the vertices of P raised to the power equal to the degree of the density function. Using this, we solve the inverse moment problem for the set of, not necessarily convex, polytopes having a given set S of vertices. Under a weak non-degeneracy assumption we also show that the uniform measure supported on any such polytope is a linear combination of uniform measures supported on simplices with vertices in S.

math.MG↗
arXiv2016

Tight Lower Bounds for Multiplicative Weights Algorithmic Families

We study the fundamental problem of prediction with expert advice and develop regret lower bounds for a large family of algorithms for this problem. We develop simple adversarial primitives, that lend themselves to various combinations leading to sharp lower bounds for many algorithmic families. We use these primitives to show that the classic Multiplicative Weights Algorithm (MWA) has a regret of $\sqrt{\frac{T \ln k}{2}}$, there by completely closing the gap between upper and lower bounds. We further show a regret lower bound of $\frac{2}{3}\sqrt{\frac{T\ln k}{2}}$ for a much more general family of algorithms than MWA, where the learning rate can be arbitrarily varied over time, or even picked from arbitrary distributions over time. We also use our primitives to construct adversaries in the geometric horizon setting for MWA to precisely characterize the regret at $\frac{0.391}{\sqrtδ}$ for the case of $2$ experts and a lower bound of $\frac{1}{2}\sqrt{\frac{\ln k}{2δ}}$ for the case of arbitrary number of experts $k$.

cs.LG↗
arXiv2016

Towards Optimal Algorithms for Prediction with Expert Advice

We study the classical problem of prediction with expert advice in the adversarial setting with a geometric stopping time. In 1965, Cover gave the optimal algorithm for the case of 2 experts. In this paper, we design the optimal algorithm, adversary and regret for the case of 3 experts. Further, we show that the optimal algorithm for $2$ and $3$ experts is a probability matching algorithm (analogous to Thompson sampling) against a particular randomized adversary. Remarkably, our proof shows that the probability matching algorithm is not only optimal against this particular randomized adversary, but also minimax optimal. Our analysis develops upper and lower bounds simultaneously, analogous to the primal-dual method. Our analysis of the optimal adversary goes through delicate asymptotics of the random walk of a particle between multiple walls. We use the connection we develop to random walks to derive an improved algorithm and regret bound for the case of $4$ experts, and, provide a general framework for designing the optimal algorithm and adversary for an arbitrary number of experts.

cs.LG↗
arXiv2016

Procrastination with variable present bias

Individuals working towards a goal often exhibit time inconsistent behavior, making plans and then failing to follow through. One well-known model of such behavioral anomalies is present-bias discounting: individuals over-weight present costs by a bias factor. This model explains many time-inconsistent behaviors, but can make stark predictions in many settings: individuals either follow the most efficient plan for reaching their goal or procrastinate indefinitely. We propose a modification in which the present-bias parameter can vary over time, drawn independently each step from a fixed distribution. Following Kleinberg and Oren (2014), we use a weighted task graph to model task planning, and measure the cost of procrastination as the relative expected cost of the chosen path versus the optimal path. We use a novel connection to optimal pricing theory to describe the structure of the worst-case task graph for any present-bias distribution. We then leverage this structure to derive conditions on the bias distribution under which the worst-case ratio is exponential (in time) or constant. We also examine conditions on the task graph that lead to improved procrastination ratios: graphs with a uniformly bounded distance to the goal, and graphs in which the distance to the goal monotonically decreases on any path.

cs.GT↗
arXiv2015

On Welfare Approximation and Stable Pricing

We study the power of item-pricing as a tool for approximately optimizing social welfare in a combinatorial market. We consider markets with $m$ indivisible items and $n$ buyers. The goal is to set prices to the items so that, when agents purchase their most demanded sets simultaneously, no conflicts arise and the obtained allocation has nearly optimal welfare. For gross substitutes valuations, it is well known that it is possible to achieve optimal welfare in this manner. We ask: can one achieve approximately efficient outcomes for valuations beyond gross substitutes? We show that even for submodular valuations, and even with only two buyers, one cannot guarantee an approximation better than $Ω(\sqrt{m})$. The same lower bound holds for the class of single-minded buyers as well. Beyond the negative results on welfare approximation, our results have daunting implications on revenue approximation for these valuation classes: in order to obtain good approximation to the collected revenue, one would necessarily need to abandon the common approach of comparing the revenue to the optimal welfare; a fundamentally new approach would be required.

cs.GT↗
arXiv2015

Liquid Price of Anarchy

Incorporating budget constraints into the analysis of auctions has become increasingly important, as they model practical settings more accurately. The social welfare function, which is the standard measure of efficiency in auctions, is inadequate for settings with budgets, since there may be a large disconnect between the value a bidder derives from obtaining an item and what can be liquidated from her. The Liquid Welfare objective function has been suggested as a natural alternative for settings with budgets. Simple auctions, like simultaneous item auctions, are evaluated by their performance at equilibrium using the Price of Anarchy (PoA) measure -- the ratio of the objective function value of the optimal outcome to the worst equilibrium. Accordingly, we evaluate the performance of simultaneous item auctions in budgeted settings by the Liquid Price of Anarchy (LPoA) measure -- the ratio of the optimal Liquid Welfare to the Liquid Welfare obtained in the worst equilibrium. Our main result is that the LPoA for mixed Nash equilibria is bounded by a constant when bidders are additive and items can be divided into sufficiently many discrete parts. Our proofs are robust, and can be extended to achieve similar bounds for simultaneous second price auctions as well as Bayesian Nash equilibria. For pure Nash equilibria, we establish tight bounds on the LPoA for the larger class of fractionally-subadditive valuations. To derive our results, we develop a new technique in which some bidders deviate (surprisingly) toward a non-optimal solution. In particular, this technique does not fit into the smoothness framework.

cs.GT↗
arXiv2014

Combinatorial Auctions via Posted Prices

We study anonymous posted price mechanisms for combinatorial auctions in a Bayesian framework. In a posted price mechanism, item prices are posted, then the consumers approach the seller sequentially in an arbitrary order, each purchasing her favorite bundle from among the unsold items at the posted prices. These mechanisms are simple, transparent and trivially dominant strategy incentive compatible (DSIC). We show that when agent preferences are fractionally subadditive (which includes all submodular functions), there always exist prices that, in expectation, obtain at least half of the optimal welfare. Our result is constructive: given black-box access to a combinatorial auction algorithm A, sample access to the prior distribution, and appropriate query access to the sampled valuations, one can compute, in polytime, prices that guarantee at least half of the expected welfare of A. As a corollary, we obtain the first polytime (in n and m) constant-factor DSIC mechanism for Bayesian submodular combinatorial auctions, given access to demand query oracles. Our results also extend to valuations with complements, where the approximation factor degrades linearly with the level of complementarity.

cs.GT↗
arXiv2014

Competitive analysis via benchmark decomposition

We propose a uniform approach for the design and analysis of prior-free competitive auctions and online auctions. Our philosophy is to view the benchmark function as a variable parameter of the model and study a broad class of functions instead of a individual target benchmark. We consider a multitude of well-studied auction settings, and improve upon a few previous results. (1) Multi-unit auctions. Given a $β$-competitive unlimited supply auction, the best previously known multi-unit auction is $2β$-competitive. We design a $(1+β)$-competitive auction reducing the ratio from $4.84$ to $3.24$. These results carry over to matroid and position auctions. (2) General downward-closed environments. We design a $6.5$-competitive auction improving upon the ratio of $7.5$. Our auction is noticeably simpler than the previous best one. (3) Unlimited supply online auctions. Our analysis yields an auction with a competitive ratio of $4.12$, which significantly narrows the margin of $[4,4.84]$ previously known for this problem. A particularly important tool in our analysis is a simple decomposition lemma, which allows us to bound the competitive ratio against a sum of benchmark functions. We use this lemma in a "divide and conquer" fashion by dividing the target benchmark into the sum of simpler functions.

cs.GT↗
arXiv2014

Short sequences of improvement moves lead to approximate equilibria in constraint satisfaction games

We present an algorithm that computes approximate pure Nash equilibria in a broad class of constraint satisfaction games that generalize the well-known cut and party affiliation games. Our results improve previous ones by Bhalgat et al.~(EC 10) in terms of the obtained approximation guarantee. More importantly, our algorithm identifies a polynomially-long sequence of improvement moves from any initial state to an approximate equilibrium in these games. The existence of such short sequences is an interesting structural property which, to the best of our knowledge, was not known before. Our techniques adapt and extend our previous work for congestion games (FOCS 11) but the current analysis is considerably simpler.

cs.GT↗
arXiv2014

Optimal Competitive Auctions

We study the design of truthful auctions for selling identical items in unlimited supply (e.g., digital goods) to n unit demand buyers. This classic problem stands out from profit-maximizing auction design literature as it requires no probabilistic assumptions on buyers' valuations and employs the framework of competitive analysis. Our objective is to optimize the worst-case performance of an auction, measured by the ratio between a given benchmark and revenue generated by the auction. We establish a sufficient and necessary condition that characterizes competitive ratios for all monotone benchmarks. The characterization identifies the worst-case distribution of instances and reveals intrinsic relations between competitive ratios and benchmarks in the competitive analysis. With the characterization at hand, we show optimal competitive auctions for two natural benchmarks. The most well-studied benchmark $\mathcal{F}^{(2)}(\cdot)$ measures the envy-free optimal revenue where at least two buyers win. Goldberg et al. [13] showed a sequence of lower bounds on the competitive ratio for each number of buyers n. They conjectured that all these bounds are tight. We show that optimal competitive auctions match these bounds. Thus, we confirm the conjecture and settle a central open problem in the design of digital goods auctions. As one more application we examine another economically meaningful benchmark, which measures the optimal revenue across all limited-supply Vickrey auctions. We identify the optimal competitive ratios to be $(\frac{n}{n-1})^{n-1}-1$ for each number of buyers n, that is $e-1$ as $n$ approaches infinity.

cs.GT↗
arXiv2013

Combinatorial Walrasian Equilibrium

We study a combinatorial market design problem, where a collection of indivisible objects is to be priced and sold to potential buyers subject to equilibrium constraints.The classic solution concept for such problems is Walrasian Equilibrium (WE), which provides a simple and transparent pricing structure that achieves optimal social welfare. The main weakness of the WE notion is that it exists only in very restrictive cases. To overcome this limitation, we introduce the notion of a Combinatorial Walrasian equilibium (CWE), a natural relaxation of WE. The difference between a CWE and a (non-combinatorial) WE is that the seller can package the items into indivisible bundles prior to sale, and the market does not necessarily clear. We show that every valuation profile admits a CWE that obtains at least half of the optimal (unconstrained) social welfare. Moreover, we devise a poly-time algorithm that, given an arbitrary allocation X, computes a CWE that achieves at least half of the welfare of X. Thus, the economic problem of finding a CWE with high social welfare reduces to the algorithmic problem of social-welfare approximation. In addition, we show that every valuation profile admits a CWE that extracts a logarithmic fraction of the optimal welfare as revenue. Finally, these results are complemented by strong lower bounds when the seller is restricted to using item prices only, which motivates the use of bundles. The strength of our results derives partly from their generality - our results hold for arbitrary valuations that may exhibit complex combinations of substitutes and complements.

cs.GT↗
arXiv2012

Convex curves and a Poisson imitation of lattices

We solve a randomized version of the following open question: is there a strictly convex, bounded curve γin the plane such that the number of rational points on γ, with denominator $n$, approaches infinity with $n$? Although this natural problem appears to be out of reach using current methods, we consider a probabilistic analogue using a spatial Poisson-process that simulates the refined rational lattice $\frac{1}{d} Z^2$, which we call $M_d$, for each natural number $d$. The main result here is that with probability 1 there exists a strictly convex, bounded curve γsuch that the number of spatial Poisson points on γ, with intensity $d$, approaches infinity with $d$. The methods include the notion of a generalized affine length of a convex curve, defined by Petrov (2007).

math.MG↗
arXiv2012

Simultaneous Auctions are (almost) Efficient

Simultaneous item auctions are simple procedures for allocating items to bidders with potentially complex preferences over different item sets. In a simultaneous auction, every bidder submits bids on all items simultaneously. The allocation and prices are then resolved for each item separately, based solely on the bids submitted on that item. Such procedures occur in practice (e.g. eBay) but are not truthful. We study the efficiency of Bayesian Nash equilibrium (BNE) outcomes of simultaneous first- and second-price auctions when bidders have complement-free (a.k.a. subadditive) valuations. We show that the expected social welfare of any BNE is at least 1/2 of the optimal social welfare in the case of first-price auctions, and at least 1/4 in the case of second-price auctions. These results improve upon the previously-known logarithmic bounds, which were established by [Hassidim, Kaplan, Mansour and Nisan '11] for first-price auctions and by [Bhawalkar and Roughgarden '11] for second-price auctions.

cs.GT↗
arXiv2012

Structure results for multiple tilings in 3D

We study multiple tilings of 3-dimensional Euclidean space by a convex body. In a multiple tiling, a convex body $P$ is translated with a discrete multiset $Λ$ in such a way that each point of the space gets covered exactly $k$ times, except perhaps the translated copies of the boundary of $P$. It is known that all possible multiple tilers in 3D are zonotopes. In 2D it was known by the work of M. Kolountzakis that, unless $P$ is a parallelogram, the multiset of translation vectors $Λ$ must be a finite union of translated lattices (also known as quasi periodic sets). In that work [Kolountzakis, 2002], the author asked whether the same quasi-periodic structure on the translation vectors would be true in 3D. Here we prove that this conclusion is indeed true for 3D. Namely, we show that if $P$ is a convex multiple tiler in 3D, with a discrete multiset $Λ$ of translation vectors, then $Λ$ has to be a finite union of translated lattices, unless $P$ belongs to a special class of zonotopes. This exceptional class consists of two-flat zonotopes $P$, defined by the Minkowski sum of $n+m$ line segments that lie in the union of two different two-dimensional subspaces $H_1$ and $H_2$. Equivalently, a two-flat zonotope $P$ may be thought of as the Minkowski sum of two 2-dimensional symmetric polygons one of which may degenerate into a single line segment. It turns out that rational two-flat zonotopes admit a multiple tiling with an aperiodic (non-quasi-periodic) set of translation vectors $Λ$. We note that it may be quite difficult to offer a visualization of these 3-dimensional non-quasi-periodic tilings, and that we discovered them by using Fourier methods.

math.CO↗
arXiv2012

The inverse moment problem for convex polytopes

The goal of this paper is to present a general and novel approach for the reconstruction of any convex d-dimensional polytope P, from knowledge of its moments. In particular, we show that the vertices of an N-vertex polytope in R^d can be reconstructed from the knowledge of O(DN) axial moments (w.r.t. to an unknown polynomial measure od degree D) in d+1 distinct generic directions. Our approach is based on the collection of moment formulas due to Brion, Lawrence, Khovanskii-Pukhikov, and Barvinok that arise in the discrete geometry of polytopes, and what variously known as Prony's method, or Vandermonde factorization of finite rank Hankel matrices.

math.NA↗
arXiv2012

Budget Feasible Mechanism Design via Random Sampling

Budget feasible mechanism considers algorithmic mechanism design questions where there is a budget constraint on the total payment of the mechanism. An important question in the field is that under which valuation domains there exist budget feasible mechanisms that admit `small' approximations (compared to a socially optimal solution). Singer \cite{PS10} showed that additive and submodular functions admit a constant approximation mechanism. Recently, Dobzinski, Papadimitriou, and Singer \cite{DPS11} gave an $O(\log^2n)$ approximation mechanism for subadditive functions and remarked that: "A fundamental question is whether, regardless of computational constraints, a constant-factor budget feasible mechanism exists for subadditive function." In this paper, we give the first attempt to this question. We give a polynomial time $O(\frac{\log n}{\log\log n})$ sub-logarithmic approximation ratio mechanism for subadditive functions, improving the best known ratio $O(\log^2 n)$. Further, we connect budget feasible mechanism design to the concept of approximate core in cooperative game theory, and show that there is a mechanism for subadditive functions whose approximation is, via a characterization of the integrality gap of a linear program, linear to the largest value to which an approximate core exists. Our result implies in particular that the class of XOS functions, which is a superclass of submodular functions, admits a constant approximation mechanism. We believe that our work could be a solid step towards solving the above fundamental problem eventually, and possibly, with an affirmative answer.

cs.GT↗