arXiv subjects

Paul G. Spirakis

Publications and source records attributed to Paul G. Spirakis.

At least 37 records · Page 2Linked to original sources

arXiv2022

Single MCMC Chain Parallelisation on Decision Trees

Decision trees are highly famous in machine learning and usually acquire state-of-the-art performance. Despite that, well-known variants like CART, ID3, random forest, and boosted trees miss a probabilistic version that encodes prior assumptions about tree structures and shares statistical strength between node parameters. Existing work on Bayesian decision trees depend on Markov Chain Monte Carlo (MCMC), which can be computationally slow, especially on high dimensional data and expensive proposals. In this study, we propose a method to parallelise a single MCMC decision tree chain on an average laptop or personal computer that enables us to reduce its run-time through multi-core processing while the results are statistically identical to conventional sequential implementation. We also calculate the theoretical and practical reduction in run time, which can be obtained utilising our method on multi-processor architectures. Experiments showed that we could achieve 18 times faster running time provided that the serial and the parallel implementation are statistically identical.

cs.AI↗

arXiv2022

The complexity of computing optimum labelings for temporal connectivity

A graph is temporally connected if there exists a strict temporal path, i.e. a path whose edges have strictly increasing labels, from every vertex $u$ to every other vertex $v$. In this paper we study temporal design problems for undirected temporally connected graphs. The basic setting of these optimization problems is as follows: given a connected undirected graph $G$, what is the smallest number $|λ|$ of time-labels that we need to add to the edges of $G$ such that the resulting temporal graph $(G,λ)$ is temporally connected? As it turns out, this basic problem, called MINIMUM LABELING (ML), can be optimally solved in polynomial time. However, exploiting the temporal dimension, the problem becomes more interesting and meaningful in its following variations, which we investigate in this paper. First we consider the problem MIN. AGED LABELING (MAL) of temporally connecting the graph when we are given an upper-bound on the allowed age (i.e. maximum label) of the obtained temporal graph $(G,λ)$. Second we consider the problem MIN. STEINER LABELING (MSL), where the aim is now to have a temporal path between any pair of "terminals" vertices which lie in a subset $R\subseteq V$. This relaxed problem resembles STEINER TREE in static graphs. However, due to the requirement of strictly increasing labels in a temporal path, STEINER TREE is not a special case of MSL. Finally we consider the age-restricted version of MSL, namely MIN. AGED STEINER LABELING (MASL). Our main results are threefold: we prove that (i) MAL becomes NP-complete on undirected graphs, while (ii) MASL becomes W[1]-hard with respect to the number $|R|$ of terminals. On the other hand we prove that (iii) although the age-unrestricted problem MSL is NP-hard, it is in FPT with respect to the number $|R|$ of terminals. That is, adding the age restriction, makes the above problems strictly harder.

cs.DS↗

arXiv2021

Threshold-based Network Structural Dynamics

The interest in dynamic processes on networks is steadily rising in recent years. In this paper, we consider the $(α,β)$-Thresholded Network Dynamics ($(α,β)$-Dynamics), where $α\leq β$, in which only structural dynamics (dynamics of the network) are allowed, guided by local thresholding rules executed in each node. In particular, in each discrete round $t$, each pair of nodes $u$ and $v$ that are allowed to communicate by the scheduler, computes a value $\mathcal{E}(u,v)$ (the potential of the pair) as a function of the local structure of the network at round $t$ around the two nodes. If $\mathcal{E}(u,v) < α$ then the link (if it exists) between $u$ and $v$ is removed; if $α\leq \mathcal{E}(u,v) < β$ then an existing link among $u$ and $v$ is maintained; if $β\leq \mathcal{E}(u,v)$ then a link between $u$ and $v$ is established if not already present. The microscopic structure of $(α,β)$-Dynamics appears to be simple, so that we are able to rigorously argue about it, but still flexible, so that we are able to design meaningful microscopic local rules that give rise to interesting macroscopic behaviors. Our goals are the following: a) to investigate the properties of the $(α,β)$-Thresholded Network Dynamics and b) to show that $(α,β)$-Dynamics is expressive enough to solve complex problems on networks. Our contribution in these directions is twofold. We rigorously exhibit the claim about the expressiveness of $(α,β)$-Dynamics, both by designing a simple protocol that provably computes the $k$-core of the network as well as by showing that $(α,β)$-Dynamics is in fact Turing-Complete. Second and most important, we construct general tools for proving stabilization that work for a subclass of $(α,β)$-Dynamics and prove speed of convergence in a restricted setting.

cs.DC↗

arXiv2021

Walrasian Equilibria in Markets with Small Demands

We study the complexity of finding a Walrasian equilibrium in markets where the agents have $k$-demand valuations. These valuations are an extension of unit-demand valuations where a bundle's value is the maximum of its $k$-subsets' values. For unit-demand agents, where the existence of a Walrasian equilibrium is guaranteed, we show that the problem is in quasi-NC. For $k=2$, we show that it is NP-hard to decide if a Walrasian equilibrium exists even if the valuations are fractionally subadditive (XOS), while for $k=3$ the hardness carries over to budget-additive valuations. In addition, we give a polynomial-time algorithm for markets with 2-demand single-minded valuations, or unit-demand valuations.

cs.GT↗

arXiv2021

Computing Exact Solutions of Consensus Halving and the Borsuk-Ulam Theorem

We study the problem of finding an exact solution to the consensus halving problem. While recent work has shown that the approximate version of this problem is PPA-complete, we show that the exact version is much harder. Specifically, finding a solution with $n$ cuts is FIXP-hard, and deciding whether there exists a solution with fewer than $n$ cuts is ETR-complete. We also give a QPTAS for the case where each agent's valuation is a polynomial. Along the way, we define a new complexity class BU, which captures all problems that can be reduced to solving an instance of the Borsuk-Ulam problem exactly. We show that FIXP $\subseteq$ BU $\subseteq$ TFETR and that LinearBU $=$ PPA, where LinearBU is the subclass of BU in which the Borsuk-Ulam instance is specified by a linear arithmetic circuit.

cs.CC↗

arXiv2020

Gathering in 1-Interval Connected Graphs

We examine the problem of gathering $k \geq 2$ agents (or multi-agent rendezvous) in dynamic graphs which may change in every synchronous round but remain always connected ($1$-interval connectivity) [KLO10]. The agents are identical and without explicit communication capabilities, and are initially positioned at different nodes of the graph. The problem is for the agents to gather at the same node, not fixed in advance. We first show that the problem becomes impossible to solve if the graph has a cycle. In light of this, we study a relaxed version of this problem, called weak gathering. We show that only in unicyclic graphs weak gathering is solvable, and we provide a deterministic algorithm for this problem that runs in polynomial number of rounds.

cs.DC↗

arXiv2020

Exact and Approximate Algorithms for Computing a Second Hamiltonian Cycle

In this paper we consider the following total functional problem: Given a cubic Hamiltonian graph $G$ and a Hamiltonian cycle $C_0$ of $G$, how can we compute a second Hamiltonian cycle $C_1 \neq C_0$ of $G$? Cedric Smith proved in 1946, using a non-constructive parity argument, that such a second Hamiltonian cycle always exists. Our main result is an algorithm which computes the second Hamiltonian cycle in time $O(n \cdot 2^{(0.3-\varepsilon)n})$ time, for some positive constant $\varepsilon>0$, and in polynomial space, thus improving the state of the art running time for solving this problem. Our algorithm is based on a fundamental structural property of Thomason's lollipop algorithm, which we prove here for the first time. In the direction of approximating the length of a second cycle in a Hamiltonian graph $G$ with a given Hamiltonian cycle $C_0$ (where we may not have guarantees on the existence of a second Hamiltonian cycle), we provide a linear-time algorithm computing a second cycle with length at least $n - 4α(\sqrt{n}+2α)+8$, where $α= \frac{Δ-2}{δ-2}$ and $δ,Δ$ are the minimum and the maximum degree of the graph, respectively. This approximation result also improves the state of the art.

cs.DS↗

arXiv2020

Matching in Stochastically Evolving Graphs

This paper studies the maximum cardinality matching problem in stochastically evolving graphs. We formally define the arrival-departure model with stochastic departures. There, a graph is sampled from a specific probability distribution and it is revealed as a series of snapshots. Our goal is to study algorithms that create a large matching in the sampled graphs. We define the price of stochasticity for this problem which intuitively captures the loss of any algorithm in the worst case in the size of the matching due to the uncertainty of the model. Furthermore, we prove the existence of a deterministic optimal algorithm for the problem. In our second set of results we show that we can efficiently approximate the expected size of a maximum cardinality matching by deriving a fully randomized approximation scheme (FPRAS) for it. The FPRAS is the backbone of a probabilistic algorithm that is optimal when the model is defined over two timesteps. Our last result is an upper bound of $\frac{2}{3}$ on the price of stochasticity. This means that there is no algorithm that can match more than $\frac{2}{3}$ of the edges of an optimal matching in hindsight.

cs.DS↗

arXiv2020

Crystal Structure Prediction via Oblivious Local Search

We study Crystal Structure Prediction, one of the major problems in computational chemistry. This is essentially a continuous optimization problem, where many different, simple and sophisticated, methods have been proposed and applied. The simple searching techniques are easy to understand, usually easy to implement, but they can be slow in practice. On the other hand, the more sophisticated approaches perform well in general, however almost all of them have a large number of parameters that require fine tuning and, in the majority of the cases, chemical expertise is needed in order to properly set them up. In addition, due to the chemical expertise involved in the parameter-tuning, these approaches can be {\em biased} towards previously-known crystal structures. Our contribution is twofold. Firstly, we formalize the Crystal Structure Prediction problem, alongside several other intermediate problems, from a theoretical computer science perspective. Secondly, we propose an oblivious algorithm for Crystal Structure Prediction that is based on local search. Oblivious means that our algorithm requires minimal knowledge about the composition we are trying to compute a crystal structure for. In addition, our algorithm can be used as an intermediate step by {\em any} method. Our experiments show that our algorithms outperform the standard basin hopping, a well studied algorithm for the problem.

cs.CE↗

arXiv2020

Approximating the Existential Theory of the Reals

The Existential Theory of the Reals (ETR) consists of existentially quantified Boolean formulas over equalities and inequalities of polynomial functions of variables in $\mathbb{R}$. In this paper we propose and study the approximate existential theory of the reals ($ε$-ETR), in which the constraints only need to be satisfied approximately. We first show that when the domain of the variables is $\mathbb{R}$ then $ε$-ETR = ETR under polynomial time reductions, and then study the constrained $ε$-ETR problem when the variables are constrained to lie in a given bounded convex set. Our main theorem is a sampling theorem, similar to those that have been proved for approximate equilibria in normal form games. It discretizes the domain in a grid-like manner whose density depends on various properties of the formula. A consequence of our theorem is that we obtain a quasi-polynomial time approximation scheme (QPTAS) for a fragment of constrained $ε$-ETR. We use our theorem to create several new PTAS and QPTAS algorithms for problems from a variety of fields.

cs.CC↗

arXiv2020

Distributed Computation and Reconfiguration in Actively Dynamic Networks

In this paper, we study systems of distributed entities that can actively modify their communication network. This gives rise to distributed algorithms that apart from communication can also exploit network reconfiguration in order to carry out a given task. At the same time, the distributed task itself may now require global reconfiguration from a given initial network $G_s$ to a target network $G_f$ from a family of networks having some good properties, like small diameter. With reasonably powerful computational entities, there is a straightforward algorithm that transforms any $G_s$ into a spanning clique in $O(\log n)$ time. The algorithm can then compute any global function on inputs and reconfigure to any target network in one round. We argue that such a strategy is impractical for real applications. In real dynamic networks there is a cost associated with creating and maintaining connections. To formally capture such costs, we define three edge-complexity measures: the \emph{total edge activations}, the \emph{maximum activated edges per round}, and the \emph{maximum activated degree of a node}. The clique formation strategy highlighted above, maximizes all of them. We aim at improved algorithms that achieve (poly)log$(n)$ time while minimizing the edge-complexity for the general task of transforming any $G_s$ into a $G_f$ of diameter (poly)log$(n)$. We give three distributed algorithms. The first runs in $O(\log n)$ time, with at most $2n$ active edges per round, an optimal total of $O(n\log n)$ edge activations, a maximum degree $n-1$, and a target network of diameter 2. The second achieves bounded degree by paying an additional logarithmic factor in time and in total edge activations and gives a target network of diameter $O(\log n)$. Our third algorithm shows that if we slightly increase the maximum degree to polylog$(n)$ then we can achieve a running time of $o(\log^2 n)$.

cs.DC↗

arXiv2019

The Price of Stability of Weighted Congestion Games

We give exponential lower bounds on the Price of Stability (PoS) of weighted congestion games with polynomial cost functions. In particular, for any positive integer $d$ we construct rather simple games with cost functions of degree at most $d$ which have a PoS of at least $\varOmega(Φ_d)^{d+1}$, where $Φ_d\sim d/\ln d$ is the unique positive root of equation $x^{d+1}=(x+1)^d$. This almost closes the huge gap between $\varTheta(d)$ and $Φ_d^{d+1}$. Our bound extends also to network congestion games. We further show that the PoS remains exponential even for singleton games. More generally, we provide a lower bound of $\varOmega((1+1/α)^d/d)$ on the PoS of $α$-approximate Nash equilibria for singleton games. All our lower bounds hold for mixed and correlated equilibria as well. On the positive side, we give a general upper bound on the PoS of $α$-approximate Nash equilibria, which is sensitive to the range $W$ of the player weights and the approximation parameter $α$. We do this by explicitly constructing a novel approximate potential function, based on Faulhaber's formula, that generalizes Rosenthal's potential in a continuous, analytic way. From the general theorem, we deduce two interesting corollaries. First, we derive the existence of an approximate pure Nash equilibrium with PoS at most $(d+3)/2$; the equilibrium's approximation parameter ranges from $\varTheta(1)$ to $d+1$ in a smooth way with respect to $W$. Secondly, we show that for unweighted congestion games, the PoS of $α$-approximate Nash equilibria is at most $(d+1)/α$.

cs.GT↗

arXiv2019

Temporal Vertex Cover with a Sliding Time Window

Modern, inherently dynamic systems are usually characterized by a network structure, i.e. an underlying graph topology, which is subject to discrete changes over time. Given a static underlying graph $G$, a temporal graph can be represented via an assignment of a set of integer time-labels to every edge of $G$, indicating the discrete time steps when this edge is active. While most of the recent theoretical research on temporal graphs has focused on the notion of a temporal path and other "path-related" temporal notions, only few attempts have been made to investigate "non-path" temporal graph problems. In this paper, motivated by applications in sensor and in transportation networks, we introduce and study two natural temporal extensions of the classical problem Vertex Cover. In both cases we wish to minimize the total number of "vertex appearances" that are needed to "cover" the whole temporal graph. In our first problem, Temporal Vertex Cover, the aim is to cover every edge at least once during the lifetime of the temporal graph, where an edge can be covered by one of its endpoints, only at a time step when it is active. In our second, more pragmatic variation Sliding Window Temporal Vertex Cover, we are also given a natural number $Δ$, and our aim is to cover every edge at least once at every $Δ$ consecutive time steps. We present a thorough investigation of the computational complexity and approximability of these two temporal covering problems. In particular, we provide strong hardness results, complemented by various approximation and exact algorithms. Some of our algorithms are polynomial-time, while others are asymptotically almost optimal under the Exponential Time Hypothesis (ETH) and other plausible complexity assumptions.

cs.CC↗

arXiv2019

Connected Subgraph Defense Games

We study a security game over a network played between a $defender$ and $k$ $attackers$. Every attacker chooses, probabilistically, a node of the network to damage. The defender chooses, probabilistically as well, a connected induced subgraph of the network of $λ$ nodes to scan and clean.Each attacker wishes to maximize the probability of escaping her cleaning by the defender. On the other hand, the goal of the defender is to maximize the expected number of attackers that she catches. This game is a generalization of the model from the seminal paper of Mavronicolas et al. "The price of defense" (MFCS'06). We are interested in Nash equilibria (NE) of this game, as well as in characterizing $defense$-$optimal$ networks which allow for the best $equilibrium$ $defense$ $ratio$, termed Price of Defense; this is the ratio of $k$ over the expected number of attackers that the defender catches in a NE. We provide characterizations of the NEs of this game and defense-optimal networks. This allows us to show that the NEs of the game coincide independently from the coordination or not of the attackers. In addition, we give an algorithm for computing NEs. Our algorithm requires exponential time in the worst case, but it is polynomial-time for $λ$ constantly close to 1 or $n$. For the special case of tree-networks, we refine our characterization which allows us to derive a polynomial-time algorithm for deciding whether a tree is defense-optimal and if this is the case it computes a defense-optimal NE. On the other hand, we prove that it is NP-hard to find a best-defense strategy if the tree is not defense-optimal. We complement this negative result with a polynomial-time constant-approximation algorithm that computes solutions that are close to optimal ones for general graphs. Finally, we provide asymptotically (almost) tight bounds for the Price of Defense for any $λ$.

cs.GT↗

arXiv2019

Fault Tolerant Network Constructors

In this work, we consider adversarial crash faults of nodes in the network constructors model $[$Michail and Spirakis, 2016$]$. We first show that, without further assumptions, the class of graph languages that can be (stably) constructed under crash faults is non-empty but small. In particular, if an unbounded number of crash faults may occur, we prove that (i) the only constructible graph language is that of spanning cliques and (ii) a strong impossibility result holds even if the size of the graphs that the protocol outputs in populations of size $n$ need only grow with $n$ (the remaining nodes being waste). When there is a finite upper bound $f$ on the number of faults, we show that it is impossible to construct any non-hereditary graph language. On the positive side, by relaxing our requirements we prove that: (i) permitting linear waste enables to construct on $n/(2f)-f$ nodes, any graph language that is constructible in the fault-free case, (ii) partial constructibility (i.e. not having to generate all graphs in the language) allows the construction of a large class of graph languages. We then extend the original model with a minimal form of fault notifications. Our main result here is a fault-tolerant universal constructor: We develop a fault-tolerant protocol for spanning line and use it to simulate a linear-space Turing Machine $M$. This allows a fault-tolerant construction of any graph accepted by $M$ in linear space, with waste $min\{n/2+f(n),\; n\}$, where $f(n)$ is the number of faults in the execution. We then prove that increasing the permissible waste to $min\{2n/3+f(n),\; n\}$ allows the construction of graphs accepted by an $O(n^2)$-space Turing Machine, which is asymptotically the maximum simulation space that we can hope for in this model. Finally, we show that logarithmic local memories can be exploited for a no-waste fault-tolerant simulation of any such protocol.

cs.DC↗

arXiv2019

How fast can we reach a target vertex in stochastic temporal graphs?

Temporal graphs are used to abstractly model real-life networks that are inherently dynamic in nature. Given a static underlying graph $G=(V,E)$, a temporal graph on $G$ is a sequence of snapshots $G_t$, one for each time step $t\geq 1$. In this paper we study stochastic temporal graphs, i.e. stochastic processes $\mathcal{G}$ whose random variables are the snapshots of a temporal graph on $G$. A natural feature observed in various real-life scenarios is a memory effect in the appearance probabilities of particular edges; i.e. the probability an edge $e\in E$ appears at time step $t$ depends on its appearance (or absence) at the previous $k$ steps. In this paper we study the hierarchy of models memory-$k$, addressing this memory effect in an edge-centric network evolution: every edge of $G$ has its own independent probability distribution for its appearance over time. Clearly, for every $k\geq 1$, memory-$(k-1)$ is a special case of memory-$k$. We make a clear distinction between the values $k=0$ ("no memory") and $k\geq 1$ ("some memory"), as in some cases these models exhibit a fundamentally different computational behavior, as our results indicate. For every $k\geq 0$ we investigate the complexity of two naturally related, but fundamentally different, temporal path (journey) problems: MINIMUM ARRIVAL and BEST POLICY. In the first problem we are looking for the expected arrival time of a foremost journey between two designated vertices $s,y$. In the second one we are looking for the arrival time of the best policy for actually choosing a particular $s$-$y$ journey. We present a detailed investigation of the computational landscape of both problems for the different values of memory $k$. Among other results we prove that, surprisingly, MINIMUM ARRIVAL is strictly harder than BEST POLICY; in fact, for $k=0$, MINIMUM ARRIVAL is #P-hard while BEST POLICY is solvable in $O(n^2)$ time.

cs.CC↗

arXiv2018

Strategic Contention Resolution in Multiple Channels

We consider the problem of resolving contention in communication networks with selfish users. In a \textit{contention game} each of $n \geq 2$ identical players has a single information packet that she wants to transmit using one of $k \geq 1$ multiple-access channels. To do that, a player chooses a slotted-time protocol that prescribes the probabilities with which at a given time-step she will attempt transmission at each channel. If more than one players try to transmit over the same channel (collision) then no transmission happens on that channel. Each player tries to minimize her own expected \textit{latency}, i.e. her expected time until successful transmission, by choosing her protocol. The natural problem that arises in such a setting is, given $n$ and $k$, to provide the players with a common, anonymous protocol (if it exists) such that no one would unilaterally deviate from it (equilibrium protocol). All previous theoretical results on strategic contention resolution examine only the case of a single channel and show that the equilibrium protocols depend on the feedback that the communication system gives to the players. Here we present multi-channel equilibrium protocols in two main feedback classes, namely \textit{acknowledgement-based} and \textit{ternary}. In particular, we provide equilibrium characterizations for more than one channels, and give specific anonymous, equilibrium protocols with finite and infinite expected latency. In the equilibrium protocols with infinite expected latency, all players transmit successfully in optimal time, i.e. $Θ(n/k)$, with probability tending to 1 as $n/k \to \infty$.

cs.GT↗

arXiv2018

Binary Search in Graphs Revisited

In the classical binary search in a path the aim is to detect an unknown target by asking as few queries as possible, where each query reveals the direction to the target. This binary search algorithm has been recently extended by [Emamjomeh-Zadeh et al., STOC, 2016] to the problem of detecting a target in an arbitrary graph. Similarly to the classical case in the path, the algorithm of Emamjomeh-Zadeh et al. maintains a candidates' set for the target, while each query asks an appropriately chosen vertex-- the "median"--which minimises a potential $Φ$ among the vertices of the candidates' set. In this paper we address three open questions posed by Emamjomeh-Zadeh et al., namely (a) detecting a target when the query response is a direction to an approximately shortest path to the target, (b) detecting a target when querying a vertex that is an approximate median of the current candidates' set (instead of an exact one), and (c) detecting multiple targets, for which to the best of our knowledge no progress has been made so far. We resolve questions (a) and (b) by providing appropriate upper and lower bounds, as well as a new potential $Γ$ that guarantees efficient target detection even by querying an approximate median each time. With respect to (c), we initiate a systematic study for detecting two targets in graphs and we identify sufficient conditions on the queries that allow for strong (linear) lower bounds and strong (polylogarithmic) upper bounds for the number of queries. All of our positive results can be derived using our new potential $Γ$ that allows querying approximate medians.

cs.DS↗