Search Constraints
Filtering by:
Language
English
Remove constraint Language: English
Resource Type
Conference Proceeding
Remove constraint Resource Type: Conference Proceeding
« Previous 
1  50 of 118

Next »
Number of results to display per page
Search Results

 Resource Type:
 Conference Proceeding
 Creator:
 Oommen, B. John, Zhan, Justin, and Crisostomo, Johanna
 Abstract:
 Anomaly detection involves identifying observations that deviate from the normal behavior of a system. One of the ways to achieve this is by identifying the phenomena that characterize "normal" observations. Subsequently, based on the characteristics of data learned from the normal observations, new observations are classified as being either normal or not. Most stateoftheart approaches, especially those which belong to the family parameterized statistical schemes, work under the assumption that the underlying distributions of the observations are stationary. That is, they assume that the distributions that are learned during the training (or learning) phase, though unknown, are not timevarying. They further assume that the same distributions are relevant even as new observations are encountered. Although such a " stationarity" assumption is relevant for many applications, there are some anomaly detection problems where stationarity cannot be assumed. For example, in network monitoring, the patterns which are learned to represent normal behavior may change over time due to several factors such as network infrastructure expansion, new services, growth of user population, etc. Similarly, in meteorology, identifying anomalous temperature patterns involves taking into account seasonal changes of normal observations. Detecting anomalies or outliers under these circumstances introduces several challenges. Indeed, the ability to adapt to changes in nonstationary environments is necessary so that anomalous observations can be identified even with changes in what would otherwise be classified as normal behavior. In this paper, we proposed to apply weak estimation theory for anomaly detection in dynamic environments. In particular, we apply this theory to detect anomaly activities in system calls. Our experimental results demonstrate that our proposal is both feasible and effective for the detection of such anomalous activities.
 Date Created:
 20120922

 Resource Type:
 Conference Proceeding
 Creator:
 Peng, Mengfei, Shi, Wei, Croft, William Lee, and Corriveau, JeanPierre
 Abstract:
 New threats to networks are constantly arising. This justifies protecting network assets and mitigating the risk associated with attacks. In a distributed environment, researchers aim, in particular, at eliminating faulty network entities. More specifically, much research has been conducted on locating a single static black hole, which is defined as a network site whose existence is known a priori and that disposes of any incoming data without leaving any trace of this occurrence. However, the prevalence of faulty nodes requires an algorithm able to (a) identify faulty nodes that can be repaired without human intervention and (b) locate black holes, which are taken to be faulty nodes whose repair does require human intervention. In this paper, we consider a specific attack model that involves multiple faulty nodes that can be repaired by mobile software agents, as well as a virus v that can infect a previously repaired faulty node and turn it into a black hole. We refer to the task of repairing multiple faulty nodes and pointing out the location of the black hole as the Faulty Node Repair and Dynamically Spawned Black Hole Search. Wefirst analyze the attack model we put forth. We then explain (a) how to identify whether a node is either (1) a normal node or (2) a repairable faulty node or (3) the black hole that has been infected by virus v during the search/repair process and, (b) how to perform the correct relevant actions. These two steps constitute a complex task, which, we explain, significantly differs from the traditional Black Hole Search. We continue by proposing an algorithm to solve this problem in an asynchronous ring network with only one whiteboard (which resides in a node called the homebase). We prove the correctness of our solution and analyze its complexity by both theoretical analysis and experiment evaluation. We conclude that, using our proposed algorithm, b + 4 agents can repair all faulty nodes and locate the black hole infected by a virus v within finite time. Our algorithm works even when the number of faulty nodes b is unknown a priori.
 Date Created:
 20170101

 Resource Type:
 Conference Proceeding
 Creator:
 Guo, Yuhong and Li, Xin
 Abstract:
 Semantic scene classification is a challenging problem in computer vision. In this paper, we present a novel multilevel active learning approach to reduce the human annotation effort for training robust scene classification models. Different from most existing active learning methods that can only query labels for selected instances at the target categorization level, i.e., the scene class level, our approach establishes a semantic framework that predicts scene labels based on a latent objectbased semantic representation of images, and is capable to query labels at two different levels, the target scene class level (abstractive high level) and the latent object class level (semantic middle level). Specifically, we develop an adaptive active learning strategy to perform multilevel label query, which maintains the default label query at the target scene class level, but switches to the latent object class level whenever an "unexpected" target class label is returned by the labeler. We conduct experiments on two standard scene classification datasets to investigate the efficacy of the proposed approach. Our empirical results show the proposed adaptive multilevel active learning approach can outperform both baseline active learning methods and a stateoftheart multilevel active learning method.
 Date Created:
 20140101

 Resource Type:
 Conference Proceeding
 Creator:
 Oommen, B. John and Polk, Spencer
 Abstract:
 The field of game playing is a particularly wellstudied area within the context of AI, leading to the development of powerful techniques, such as the alphabeta search, capable of achieving competitive game play against an intelligent opponent. It is well known that tree pruning strategies, such as alphabeta, benefit strongly from proper move ordering, that is, searching the best element first. Inspired by the formerly unrelated field of Adaptive Data Structures (ADSs), we have previously introduced the HistoryADS technique, which employs an adaptive list to achieve effective and dynamic move ordering, in a domain independent fashion, and found that it performs well in a wide range of cases. However, previous work did not compare the performance of the HistoryADS heuristic to any established move ordering strategy. In an attempt to address this problem, we present here a comparison to two wellknown, acclaimed strategies, which operate on a similar philosophy to the HistoryADS, the History Heuristic, and the Killer Moves technique. We find that, in a wide range of twoplayer and multiplayer games, at various points in the game’s progression, the HistoryADS performs at least as well as these strategies, and, in fact, outperforms them in the majority of cases.
 Date Created:
 20160101

 Resource Type:
 Conference Proceeding
 Creator:
 Oommen, B. John and Astudillo, César A.
 Abstract:
 We present a method that employs a treebased Neural Network (NN) for performing classification. The novel mechanism, apart from incorporating the information provided by unlabeled and labeled instances, rearranges the nodes of the tree as per the laws of Adaptive Data Structures (ADSs). Particularly, we investigate the Pattern Recognition (PR) capabilities of the TreeBased TopologyOriented SOM (TTOSOM) when Conditional Rotations (CONROT) [8] are incorporated into the learning scheme. The learning methodology inherits all the properties of the TTOSOMbased classifier designed in [4]. However, we now augment it with the property that frequently accessed nodes are moved closer to the root of the tree. Our experimental results show that on average, the classification capabilities of our proposed strategy are reasonably comparable to those obtained by some of the stateoftheart classification schemes that only use labeled instances during the training phase. The experiments also show that improved levels of accuracy can be obtained by imposing trees with a larger number of nodes.
 Date Created:
 20150101

 Resource Type:
 Conference Proceeding
 Creator:
 Yazidi, Anis, Oommen, B. John, and Hammer, Hugo Lewi
 Abstract:
 The problem of clustering, or unsupervised classification, has been solved by a myriad of techniques, all of which depend, either directly or implicitly, on the Bayesian principle of optimal classification. To be more specific, within a Bayesian paradigm, if one is to compare the testing sample with only a single point in the feature space from each class, the optimal Bayesian strategy would be to achieve this based on the distance from the corresponding means or central points in the respective distributions. When this principle is applied in clustering, one would assign an unassigned sample into the cluster whose mean is the closest, and this can be done in either a bottomup or a topdown manner. This paper pioneers a clustering achieved in an “AntiBayesian” manner, and is based on the breakthrough classification paradigm pioneered by Oommen et al. The latter relies on a radically different approach for classifying data points based on the noncentral quantiles of the distributions. Surprisingly and counterintuitively, this turns out to work equally or closetoequally well to an optimal supervised Bayesian scheme, which thus begs the natural extension to the unexplored arena of clustering. Our algorithm can be seen as the AntiBayesian counterpart of the wellknown kmeans algorithm (The fundamental AntiBayesian paradigm need not just be used to the kmeans principle. Rather, we hypothesize that it can be adapted to any of the scores of techniques that is indirectly based on the Bayesian paradigm.), where we assign points to clusters using quantiles rather than the clusters’ centroids. Extensive experimentation (This paper contains the prima facie results of experiments done on one and twodimensional data. The extensions to multidimensional data are not included in the interest of space, and would use the corresponding multidimensional AntiNa¨ıveBayes classification rules given in [1].) demonstrates that our AntiBayesian clustering converges fast and with precision results competitive to a kmeans clustering.
 Date Created:
 20150101

 Resource Type:
 Conference Proceeding
 Creator:
 Tavasoli, Hanane, Oommen, B. John, and Yazidi, Anis
 Abstract:
 In this paper, we propose a novel online classifier for complex data streams which are generated from nonstationary stochastic properties. Instead of using a single training model and counters to keep important data statistics, the introduced online classifier scheme provides a realtime selfadjusting learning model. The learning model utilizes the multiplicationbased update algorithm of the Stochastic Learning Weak Estimator (SLWE) at each time instant as a new labeled instance arrives. In this way, the data statistics are updated every time a new element is inserted, without requiring that we have to rebuild its model when changes occur in the data distributions. Finally, and most importantly, the model operates with the understanding that the correct classes of previouslyclassified patterns become available at a later juncture subsequent to some time instances, thus requiring us to update the training set and the training model. The results obtained from rigorous empirical analysis on multinomial distributions, is remarkable. Indeed, it demonstrates the applicability of our method on synthetic datasets, and proves the advantages of the introduced scheme.
 Date Created:
 20160101

 Resource Type:
 Conference Proceeding
 Creator:
 Polk, Spencer and Oommen, B. John
 Abstract:
 This paper pioneers the avenue of enhancing a wellknown paradigm in game playing, namely the use of Historybased heuristics, with a totallyunrelated area of computer science, the field of Adaptive Data Structures (ADSs). It is a wellknown fact that highlyregarded game playing strategies, such as alphabeta search, benefit strongly from proper move ordering, and from this perspective, the History heuristic is, probably, one of the most acclaimed techniques used to achieve AIbased game playing. Recently, the authors of this present paper have shown that techniques derived from the field of ADSs, which are concerned with query optimization in a data structure, can be applied to move ordering in multiplayer games. This was accomplished by ranking opponent threat levels. The work presented in this paper seeks to extend the utility of ADSbased techniques to twoplayer and multiplayer games, through the development of a new move ordering strategy that incorporates the historical advantages of the moves. The resultant technique, the HistoryADS heuristic, has been found to produce substantial (i.e, even up to 70%) savings in a variety of twoplayer and multiplayer games, at varying ply depths, and at both initial and midgame board states. As far as we know, results of this nature have not been reported in the literature before.
 Date Created:
 20150101

 Resource Type:
 Conference Proceeding
 Creator:
 Labiche, Yvan and Barros, Márcio
 Date Created:
 20150101

 Resource Type:
 Conference Proceeding
 Creator:
 Oommen, B. John and Kim, SangWoon
 Abstract:
 This paper deals with the relatively new field of sequencebased estimation which involves utilizing both the information in the observations and in their sequence of appearance. Our intention is to obtain Maximum Likelihood estimates by “extracting” the information contained in the observations when perceived as a sequence rather than as a set. The results of [15] introduced the concepts of Sequence Based Estimation (SBE) for the Binomial distribution. This current paper generalizes these results for the multinomial “twoatatime” scenario. We invoke a novel phenomenon called “Occlusion” that can be described as follows: By “concealing” certain observations, we map the estimation problem onto a lowerdimensional binomial space. Once these occluded SBEs have been computed, we demonstrate how the overall Multinomial SBE (MSBE) can be obtained by mapping several lowerdimensional estimates onto the original higherdimensional space. We formally prove and experimentally demonstrate the convergence of the corresponding estimates.
 Date Created:
 20160101

 Resource Type:
 Conference Proceeding
 Creator:
 Maheshwari, Anil, Nandy, Ayan, Smid, Michiel, and Das, Sandip
 Abstract:
 Consider a line segment R consisting of n facilities. Each facility is a point on R and it needs to be assigned exactly one of the colors from a given palette of c colors. At an instant of time only the facilities of one particular color are 'active' and all other facilities are 'dormant'. For the set of facilities of a particular color, we compute the one dimensional Voronoi diagram, and find the cell, i.e, a segment of maximum length. The users are assumed to be uniformly distributed over R and they travel to the nearest among the facilities of that particular color that is active. Our objective is to assign colors to the facilities in such a way that the length of the longest cell is minimized. We solve this optimization problem for various values of n and c. We propose an optimal coloring scheme for the number of facilities n being a multiple of c as well as for the general case where n is not a multiple of c. When n is a multiple of c, we compute an optimal scheme in Θ(n) time. For the general case, we propose a coloring scheme that returns the optimal in O(n2logn) time.
 Date Created:
 20140101

 Resource Type:
 Conference Proceeding
 Creator:
 Kim, SangWoon and Oommen, B. John
 Abstract:
 The Maximum Likelihood (ML) and Bayesian estimation paradigms work within the model that the data, from which the parameters are to be estimated, is treated as a set rather than as a sequence. The pioneering paper that dealt with the field of sequencebased estimation [2] involved utilizing both the information in the observations and in their sequence of appearance. The results of [2] introduced the concepts of Sequence Based Estimation (SBE) for the Binomial distribution, where the authors derived the corresponding MLE results when the samples are taken twoatatime, and then extended these for the cases when they are processed threeatatime, fouratatime etc. These results were generalized for the multinomial “twoatatime” scenario in [3]. This paper (This paper is dedicated to the memory of Dr. Mohamed Kamel, who was a close friend of the first author.) now further generalizes the results found in [3] for the multinomial case and for subsequences of length 3. The strategy used in [3] (and also here) involves a novel phenomenon called “Occlusion” that has not been reported in the field of estimation. The phenomenon can be described as follows: By occluding (hiding or concealing) certain observations, we map the estimation problem onto a lowerdimensional space, i.e., onto a binomial space. Once these occluded SBEs have been computed, the overall Multinomial SBE (MSBE) can be obtained by combining these lowerdimensional estimates. In each case, we formally prove and experimentally demonstrate the convergence of the corresponding estimates.
 Date Created:
 20160101

 Resource Type:
 Conference Proceeding
 Creator:
 Bertossi, Leopoldo
 Abstract:
 A correspondence between database tuples as causes for query answers in databases and tuplebased repairs of inconsistent databases with respect to denial constraints has already been established. In this work, answerset programs that specify repairs of databases are used as a basis for solving computational and reasoning problems about causes. Here, causes are also introduced at the attribute level by appealing to a both nullbased and attributebased repair semantics. The corresponding repair programs are presented, and they are used as a basis for computation and reasoning about attributelevel causes.
 Date Created:
 20180101

 Resource Type:
 Conference Proceeding
 Creator:
 Dujmović, Vida, De Carufel, JeanLou, Bose, Prosenjit, and Paradis, Frédérik
 Abstract:
 The wellseparated pair decomposition (WSPD) of the complete Euclidean graph defined on points in ℝ2 (Callahan and Kosaraju [JACM, 42 (1): 6790, 1995]) is a technique for partitioning the edges of the complete graph based on length into a linear number of sets. Among the many different applications of WSPDs, Callahan and Kosaraju proved that the sparse subgraph that results by selecting an arbitrary edge from each set (called WSPDspanner) is a 1 + 8/(s − 4)spanner, where s > 4 is the separation ratio used for partitioning the edges. Although competitive localrouting strategies exist for various spanners such as Yaographs, Θgraphs, and variants of Delaunay graphs, few localrouting strategies are known for any WSPDspanner. Our main contribution is a localrouting algorithm with a nearoptimal competitive routing ratio of 1 + O(1/s) on a WSPDspanner. Specifically, we present a 2local and a 1local routing algorithm on a WSPDspanner with competitive routing ratios of 1+6/(s−2)+4/s and 1+6/(s−2)+ 6/s + 4/(s2 − 2s) + 8/s2respectively.
 Date Created:
 20170101

 Resource Type:
 Conference Proceeding
 Creator:
 Lanthier, Mark, Velazquez, Elio, and Santoro, Nicola
 Abstract:
 This paper proposes a proactive solution to the Frugal Feeding Problem (FFP) in Wireless Sensor Networks. The FFP attempts to find energyefficient routes for a mobile service entity to rendezvous with each member of a team of mobile robots. Although the complexity of the FFP is similar to the Traveling Salesman Problem (TSP), we propose an efficient solution, completely distributed and localized for the case of a fixed rendezvous location (i.e., service facility with limited number of docking ports) and mobile capable entities (sensors). Our proactive solution reduces the FFP to finding energyefficient routes in a dynamic Compass Directed unit Graph (CDG). The proposed CDG incorporates ideas from forward progress routing and the directionality of compass routing in an energyaware unit subgraph. Navigating the CDG guarantees that each sensor will reach the rendezvous location in a finite number of steps. The ultimate goal of our solution is to achieve energy equilibrium (i.e., no further sensor losses due to energy starvation) by optimizing the use of the shared resource (recharge station). We also examine the impact of critical parameters such as transmission range, cost of mobility and sensor knowledge in the overall performance.
 Date Created:
 20111114

 Resource Type:
 Conference Proceeding
 Creator:
 Guo, Yuhong and Li, Xin
 Abstract:
 Multilabel classification is a central problem in many application domains. In this paper, we present a novel supervised bidirectional model that learns a lowdimensional midlevel representation for multilabel classification. Unlike traditional multilabel learning methods which identify intermediate representations from either the input space or the output space but not both, the midlevel representation in our model has two complementary parts that capture intrinsic information of the input data and the output labels respectively under the autoencoder principle while augmenting each other for the target output label prediction. The resulting optimization problem can be solved efficiently using an iterative procedure with alternating steps, while closedform solutions exist for one major step. Our experiments conducted on a variety of multilabel data sets demonstrate the efficacy of the proposed bidirectional representation learning model for multilabel classification.
 Date Created:
 20140101

 Resource Type:
 Conference Proceeding
 Creator:
 White, Anthony and SalehiAbari, Amirali
 Abstract:
 Autonomous agents require trust and reputation concepts in order to identify communities of agents with which to interact reliably in ways analogous to humans. Agent societies are invariably heterogeneous, with multiple decision making policies and actions governing their behaviour. Through the introduction of naive agents, this paper shows empirically that while learning agents can identify malicious agents through direct interaction, naive agents compromise utility through their inability to discern malicious agents. Moreover, the impact of the proportion of naive agents on the society is analyzed. The paper demonstrates that there is a need for witness interaction trust to detect naive agents in addition to the need for direct interaction trust to detect malicious agents. By proposing a set of policies, the paper demonstrates how learning agents can isolate themselves from naive and malicious agents.
 Date Created:
 20100720

 Resource Type:
 Conference Proceeding
 Creator:
 Prencipe, Giuseppe, Cáceres, Edson, Chan, Albert, and Dehne, Frank
 Abstract:
 In this paper, we present parallel algorithms for the coarse grained multicomputer (CGM) and the bulk synchronous parallel computer (BSP) for solving two well known graph problems: (1) determining whether a graph G is bipartite, and (2) determining whether a bipartite graph G is convex. Our algorithms require O(log p) and O(log2 p) communication rounds, respectively, and linear sequential work per round on a CGM with p processors and N/p local memory per processor, N=G. The algorithms assume that N/ p ≥ p€ for some fixed€ > 0, which is true for all commercially available multiprocessors. Our results imply BSP algorithms with O(log p) and O(log2 p) supersteps, respectively, O(g log(p) N p) communication time, and O(log(p) N p) local computation time. Our algorithm for determining whether a bipartite graph is convex includes a novel, coarse grained parallel, version of the PQ tree data structure introduced by Booth and Lueker. Hence, our algorithm also solves, with the same time complexity as indicated above, the problem of testing the consecutiveones property for (0, 1) matrices as well as the chordal graph recognition problem. These, in turn, have numerous applications in graph theory, DNA sequence assembly, database theory, and other areas.
 Date Created:
 20000101

 Resource Type:
 Conference Proceeding
 Creator:
 Maheshwari, Anil and Zeh, Norbert
 Abstract:
 We present external memory algorithms for outerplanarity testing, embedding outerplanar graphs, breadthfirst search (BFS) and depthfirst search (DFS) in outerplanar graphs, and finding a2separator of size 2 for a given outerplanar graph. Our algorithms take O(sort(N)) I/Os and can easily be improved to take O (perm (N)) I/Os, as all these problems have linear time solutions in internal memory. For BFS, DFS, and outerplanar embedding we show matching lower bounds.
 Date Created:
 19990101

 Resource Type:
 Conference Proceeding
 Creator:
 Morin, Pat and Bose, Prosenjit
 Abstract:
 We consider online routing strategies for routing between the vertices of embedded planar straight line graphs. Our results include (1) two deterministic memoryless routing strategies, one that works for all Delaunay triangulations and the other that works for all regular triangulations, (2) a randomized memoryless strategy that works for all triangulations, (3) an O(1) memory strategy that works for all convex subdivisions, (4) an O(1) memory strategy that approximates the shortest path in Delaunay triangulations, and (5) theoretical and experimental results on the competitiveness of these strategies.
 Date Created:
 19990101

 Resource Type:
 Conference Proceeding
 Creator:
 Krizanc, Danny, Kranakis, Evangelos, and Kirousis, Lefteris M.
 Abstract:
 Let φ be a random Boolean formula that is an instance of 3SAT. We consider the problem of computing the least real number such that if the ratio of the number of clauses over the number of variables of φ strictly exceeds κ, then φ is almost certainly unsatisfiable. By a well known and more or less straightforward argument, it can be shown that κ 3.
 Date Created:
 19960101

 Resource Type:
 Conference Proceeding
 Creator:
 Maheshwari, Anil, Sack, JörgRüdiger, Lanthier, Mark, and Aleksandrov, Lyudmil
 Date Created:
 19980101

 Resource Type:
 Conference Proceeding
 Creator:
 Peleg, David, Krizanc, Danny, Kirousis, Lefteris M., Kranakis, Evangelos, Kaklamanis, Christos, and Bose, Prosenjit
 Abstract:
 In wireless communication, the signal of a typical broadcast station is transmited from a broadcast center p and reaches objects at a distance, say, R from it. In addition there is a radius r, r < R, such that the signal originating from the center of the station is so strong that human habitation within distance r from the center p should be avoided. Thus every station determines a region which is an “annulus of permissible habitation". We consider the following station layout (SL) problem: Cover a given (say, rectangular) planar region which includes a collection of orthogonal buildings with a minimum number of stations so that every point in the region is within the reach of a station, while at the same time no building is within the dangerous range of a station. We give algorithms for computing such station layouts in both the oneand twodimensional cases.
 Date Created:
 19990101

 Resource Type:
 Conference Proceeding
 Creator:
 Bose, Prosenjit and Van Renssen, André
 Abstract:
 We present tight upper and lower bounds on the spanning ratio of a large family of constrained θgraphs. We show that constrained θgraphs with 4k2 (k≥ 1 and integer) cones have a tight spanning ratio of 1+2 sin(θ/2), where θ is 2 π/ (4k+2). We also present improved upper bounds on the spanning ratio of the other families of constrained θgraphs.
 Date Created:
 20140101

 Resource Type:
 Conference Proceeding
 Creator:
 Seidel, Raimund, Dehne, Frank, and Klein, Rolf
 Abstract:
 Given a set S of s points in the plane, where do we place a new point, p, in order to maximize the area of its region in the Voronoi diagram of S and p? We study the case where the Voronoi neighbors of p are in convex position, and prove that there is at most one local maximum.
 Date Created:
 20021201

 Resource Type:
 Conference Proceeding
 Creator:
 Mannan, Mohammad, Barrera, David, Van Oorschot, Paul C., Lie, David, and Brown, Carson D.
 Abstract:
 Instead of allowing the recovery of original passwords, forgotten passwords are often reset using online mechanisms such as password verification questions (PVQ methods) and password reset links in email. These mechanisms are generally weak, exploitable, and force users to choose new passwords. Emailing the original password exposes the password to third parties. To address these issues, and to allow forgotten passwords to be securely restored, we present a scheme called Mercury. Its primary mode employs userlevel public keys and a personal mobile device (PMD) such as a smartphone, netbook, or tablet. A user generates a key pair on her PMD; the private key remains on the PMD and the public key is shared with different sites (e.g., during account setup). For password recovery, the site sends the (public key)encrypted password to the user's preregistered email address, or displays the encrypted password on a webpage, e.g., as a barcode. The encrypted password is then decrypted using the PMD and revealed to the user. A prototype implementation of Mercury is available as an Android application.
 Date Created:
 20120221

 Resource Type:
 Conference Proceeding
 Creator:
 Van Walderveen, Freek, Davoodi, Pooya, and Smid, Michiel
 Abstract:
 Given a set of n points in the plane, range diameter queries ask for the furthest pair of points in a given axisparallel rectangular range. We provide evidence for the hardness of designing spaceefficient data structures that support range diameter queries by giving a reduction from the set intersection problem. The difficulty of the latter problem is widely acknowledged and is conjectured to require nearly quadratic space in order to obtain constant query time, which is matched by known data structures for both problems, up to polylogarithmic factors. We strengthen the evidence by giving a lower bound for an important subproblem arising in solutions to the range diameter problem: computing the diameter of two convex polygons, that are separated by a vertical line and are preprocessed independently, requires almost linear time in the number of vertices of the smaller polygon, no matter how much space is used. We also show that range diameter queries can be answered much more efficiently for the case of points in convex position by describing a data structure of size O(n log n) that supports queries in O(log n) time.
 Date Created:
 20120515

 Resource Type:
 Conference Proceeding
 Creator:
 Cervera, Gimer, Barbeau, Michel, GarciaAlfaro, Joaquin, and Kranakis, Evangelos
 Abstract:
 The Hierarchical Optimized Link State Routing (HOLSR) protocol enhances the scalability and heterogeneity of traditional OLSRbased Mobile AdHoc Networks (MANETs). It organizes the network in logical levels and nodes in clusters. In every cluster, it implements the mechanisms and algorithms of the original OLSR to generate and to distribute control traffic information. However, the HOLSR protocol was designed with no security in mind. Indeed, it both inherits, from OLSR, and adds new security threats. For instance, the existence of misbehaving nodes can highly affect important HOLSR operations, such as the cluster formation. Cluster IDentification (CID) messages are implemented to organize a HOLSR network in clusters. In every message, the hop count field indicates to the receiver the distance in hops to the originator. An attacker may maliciously alter the hop count field. As a consequence, a receiver node may join a cluster head farther away than it appears. Then, the scalability properties in a HOLSR network is affected by an unbalanced distribution of nodes per cluster. We present a solution based on the use of hash chains to protect mutable fields in CID messages. As a consequence, when a misbehaving node alters the hop count field in a CID message, the receiver nodes are able of detecting and discarding the invalid message.
 Date Created:
 20120127

 Resource Type:
 Conference Proceeding
 Creator:
 Czyzowicz, Jurek, Opatrny, Jaroslav, Kranakis, Evangelos, Narayanan, Lata, Krizanc, Danny, Stacho, Ladislav, Urrutia, Jorge, Yazdani, Mohammadreza, and Lambadaris, Ioannis
 Abstract:
 A set of sensors establishes barrier coverage of a given line segment if every point of the segment is within the sensing range of a sensor. Given a line segment I, n mobile sensors in arbitrary initial positions on the line (not necessarily inside I) and the sensing ranges of the sensors, we are interested in finding final positions of sensors which establish a barrier coverage of I so that the sum of the distances traveled by all sensors from initial to final positions is minimized. It is shown that the problem is NP complete even to approximate up to constant factor when the sensors may have different sensing ranges. When the sensors have an identical sensing range we give several efficient algorithms to calculate the final destinations so that the sensors either establish a barrier coverage or maximize the coverage of the segment if complete coverage is not feasible while at the same time the sum of the distances traveled by all sensors is minimized. Some open problems are also mentioned.
 Date Created:
 20101213

 Resource Type:
 Conference Proceeding
 Creator:
 Barbeau, Michel, Kranakis, Evangelos, and GarciaAlfaro, Joaquin
 Abstract:
 The design and implementation of security threat mitigation mechanisms in RFID systems, specially in lowcost RFID tags, are gaining great attention in both industry and academia. One main focus of research interests is the authentication and privacy techniques to prevent attacks targeting the insecure wireless channel of these systems. Cryptography is a key tool to address these threats. Nevertheless, strong hardware constraints, such as production costs, power consumption, time of response, and regulations compliance, makes the use of traditional cryptography in these systems a very challenging problem. The use of lowoverhead procedures becomes the main approach to solve these challenging problems where traditional cryptography cannot fit. Recent results and trends, with an emphasis on lightweight techniques for addressing critical threats against lowcost RFID systems, are surveyed.
 Date Created:
 20100503

 Resource Type:
 Conference Proceeding
 Creator:
 Kranakis, Evangelos, Krizanc, Danny, Narayanan, Lata, and Keane, Michael
 Abstract:
 Delay (or disruption) tolerant sensor networks may be modeled as Markovian evolving graphs [1]. We present experimental evidence showing that considering multiple (possibly not shortest) paths instead of one fixed (greedy) path can decrease the expected time to deliver a packet on such a network by as much as 65 per cent depending on the probability that an edge exists in a given time interval. We provide theoretical justification for this result by studying a special case of the Markovian evolving grid graph. We analyze a natural algorithm for routing on such networks and show that it is possible to improve the expected time of delivery by up to a factor of two depending upon the probability of an edge being up during a time step and the relative positions of the source and destination. Furthermore we show that this is optimal, i.e., no other algorithm can achieve a better expected running time. As an aside, our results give high probability bounds for Knuth's toilet paper problem [11].
 Date Created:
 20091201

 Resource Type:
 Conference Proceeding
 Creator:
 He, Meng, Dillabaugh, Craig, Zeh, Norbert, and Maheshwari, Anil
 Date Created:
 20091201

 Resource Type:
 Conference Proceeding
 Creator:
 Krizanc, D., Yazdani, M., Stacho, L., Narayanan, L., Lambadaris, Ioannis, Opatrny, J., Czyzowicz, J., Kranakis, Evangelos, and Urrutia, J.
 Abstract:
 We consider n mobile sensors located on a line containing a barrier represented by a finite line segment. Sensors form a wireless sensor network and are able to move within the line. An intruder traversing the barrier can be detected only when it is within the sensing range of at least one sensor. The sensor network establishes barrier coverage of the segment if no intruder can penetrate the barrier from any direction in the plane without being detected. Starting from arbitrary initial positions of sensors on the line we are interested in finding final positions of sensors that establish barrier coverage and minimize the maximum distance traversed by any sensor. We distinguish several variants of the problem, based on (a) whether or not the sensors have identical ranges, (b) whether or not complete coverage is possible and (c) in the case when complete coverage is impossible, whether or not the maximal coverage is required to be contiguous. For the case of n sensors with identical range, when complete coverage is impossible, we give linear time optimal algorithms that achieve maximal coverage, both for the contiguous and noncontiguous case. When complete coverage is possible, we give an O(n 2) algorithm for an optimal solution, a linear time approximation scheme with approximation factor 2, and a (1∈+∈ε) PTAS. When the sensors have unequal ranges we show that a variation of the problem is NPcomplete and identify some instances which can be solved with our algorithms for sensors with unequal ranges.
 Date Created:
 20091019

 Resource Type:
 Conference Proceeding
 Creator:
 Barbeau, Michel and Laurendeau, Christine
 Abstract:
 Increasingly ubiquitous wireless technologies require novel localization techniques to pinpoint the position of an uncooperative node, whether the target be a malicious device engaging in a security exploit or a lowbattery handset in the middle of a critical emergency. Such scenarios necessitate that a radio signal source be localized by other network nodes efficiently, using minimal information. We propose two new algorithms for estimating the position of an uncooperative transmitter, based on the received signal strength (RSS) of a single target message at a set of receivers whose coordinates are known. As an extension to the concept of centroid localization, our mechanisms weigh each receiver's coordinates based on the message's relative RSS at that receiver, with respect to the span of RSS values over all receivers. The weights may decrease from the highest RSS receiver either linearly or exponentially. Our simulation results demonstrate that for all but the most sparsely populated wireless networks, our exponentially weighted mechanism localizes a target node within the regulations stipulated for emergency services location accuracy.
 Date Created:
 20090928

 Resource Type:
 Conference Proceeding
 Creator:
 Dillabaugh, Craig, He, Meng, and Maheshwari, Anil
 Abstract:
 We present two results for path traversal in trees, where the traversal is performed in an asymptotically optimal number of I/Os and the tree structure is represented succinctly. Our first result is for bottomup traversal that starts with a node in the tree T and traverses a path to the root. For blocks of size B, a tree on N nodes, and for a path of length K, we design data structures that permit traversal of the bottomup path in O(K/B) I/Os using only bits, for an arbitrarily selected constant, ε, where 0∈<∈ε<∈1. Our second result is for topdown traversal in binary trees. We store T using (3∈+∈q)N∈+∈o(N) bits, where q is the number of bits required to store a key, while topdown traversal can still be performed in an asymptotically optimal number of I/Os.
 Date Created:
 20081201

 Resource Type:
 Conference Proceeding
 Creator:
 Petriu, Dorina C. and Tawhid, Rasha
 Abstract:
 The paper proposes to integrate performance analysis in the early phases of the modeldriven development process for Software Product Lines (SPL). We start by adding generic performance annotations to the UML model representing the set of core reusable SPL assets. The annotations are generic and use the MARTE Profile recently adopted by OMG. A first model transformation realized in the Atlas Transformation Language (ATL), which is the focus of this paper, derives the UML model of a specific product with concrete MARTE performance annotations from the SPL model. A second transformation generates a Layered Queueing Network performance model for the given product by applying an existing transformation approach named PUMA, developed in previous work. The proposed technique is illustrated with an ecommerce case study that models the commonality and variability in both structural and behavioural SPL views. A product is derived and the performance of two design alternatives is compared.
 Date Created:
 20081128

 Resource Type:
 Conference Proceeding
 Creator:
 Kranakis, Evangelos and Wiese, Andreas
 Abstract:
 We investigate the problem of locally coloring and constructing special spanners of location aware Unit Disk Graphs (UDGs). First we present a local approximation algorithm for the vertex coloring problem in UDGs which uses at most four times as many colors as required by an optimal solution. Then we look at the colorability of spanners of UDGs. In particular we present a local algorithm for constructing a 4colorable spanner of a unit disk graph. The output consists of the spanner and the 4coloring. The computed spanner also has the properties that it is planar, the degree of a vertex in the spanner is at most 5 and the angles between two edges are at least π/3. By enlarging the locality distance (i.e. the size of the neighborhood which a vertex has to explore in order to compute its color) we can ensure the total weight of the spanner to be arbitrarily close to the weight of a minimum spanning tree. We prove that a local algorithm cannot compute a bipartite spanner of a unit disk graph and therefore our algorithm needs at most one color more than any local algorithm for the task requires. Moreover, we prove that there is no local algorithm for 3coloring UDGs or spanners of UDGs, even if the 3colorability of the graph (or the spanner respectively) is guaranteed in advance.
 Date Created:
 20081201

 Resource Type:
 Conference Proceeding
 Creator:
 Bose, Prosenjit, Maheshwari, Anil, Carmi, Paz, Smid, Michiel, and Farshi, Mohammad
 Abstract:
 It is wellknown that the greedy algorithm produces high quality spanners and therefore is used in several applications. However, for points in ddimensional Euclidean space, the greedy algorithm has cubic running time. In this paper we present an algorithm that computes the greedy spanner (spanner computed by the greedy algorithm) for a set of n points from a metric space with bounded doubling dimension in time using space. Since the lower bound for computing such spanners is Ω(n 2), the time complexity of our algorithm is optimal to within a logarithmic factor.
 Date Created:
 20081027

 Resource Type:
 Conference Proceeding
 Creator:
 Kranakis, Evangelos, Shi, Q., Bhattacharya, B., Wiese, A., Burmester, B., and Hu, Y.
 Abstract:
 Intrusion detection, area coverage and border surveillance are important applications of wireless sensor networks today. They can be (and are being) used to monitor large unprotected areas so as to detect intruders as they cross a border or as they penetrate a protected area. We consider the problem of how to optimally move mobile sensors to the fence (perimeter) of a region delimited by a simple polygon in order to detect intruders from either entering its interior or exiting from it. We discuss several related issues and problems, propose two models, provide algorithms and analyze their optimal mobility behavior.
 Date Created:
 20080922

 Resource Type:
 Conference Proceeding
 Creator:
 Couture, Mathieu, Smid, Michiel, Maheshwari, Anil, Bose, Prosenjit, Carmi, Paz, and Zeh, Norbert
 Abstract:
 Given an integer k ≥ 2, we consider the problem of computing the smallest real number t(k) such that for each set P of points in the plane, there exists a t(k)spanner for P that has chromatic number at most k. We prove that t(2)∈=∈3, t(3)∈=∈2, , and give upper and lower bounds on t(k) for k∈>∈4. We also show that for any ε>∈0, there exists a (1∈+∈ε)t(k)spanner for P that has O(P) edges and chromatic number at most k. Finally, we consider an online variant of the problem where the points of P are given one after another, and the color of a point must be assigned at the moment the point is given. In this setting, we prove that t(2)∈=∈3, , , and give upper and lower bounds on t(k) for k∈>∈4.
 Date Created:
 20080827

 Resource Type:
 Conference Proceeding
 Creator:
 Kranakis, Evangelos and Wiese, Andreas
 Abstract:
 We present the first local approximation schemes for maximum independent set and minimum vertex cover in unit disk graphs. In the graph model we assume that each node knows its geographic coordinates in the plane (location aware nodes). Our algorithms are local in the sense that the status of each node v (whether or not v is in the computed set) depends only on the vertices which are a constant number of hops away from v. This constant is independent of the size of the network. We give upper bounds for the constant depending on the desired approximation ratio. We show that the processing time which is necessary in order to compute the status of a single vertex is bounded by a polynomial in the number of vertices which are at most a constant number of vertices away from it. Our algorithms give the best possible approximation ratios for this setting. The technique which we use to obtain the algorithm for vertex cover can also be employed for constructing the first global PTAS for this problem in unit disk graph which does not need the embedding of the graph as part of the input.
 Date Created:
 20080701

 Resource Type:
 Conference Proceeding
 Creator:
 Biddle, Robert, Noble, James, Barr, Pippin, Fischer, Ronald, and Khaled, Rilla
 Abstract:
 Persuasive technologies are increasingly ubiquitous, but the strategies they utilise largely originate in America. Consumer behaviour research shows us that certain persuasion strategies will be more effective on some cultures than others. We claim that the existing strategies will be less effective on nonAmerican audiences than they are on American audiences, and we use information from interviews to show that there exists much scope to develop persuasive technologies from a collectivismfocused perspective. To illustrate the development of such a tool, we describe the design of a collectivismfocused financial planning tool.
 Date Created:
 20060717

 Resource Type:
 Conference Proceeding
 Creator:
 Smid, Michiel and Gudmundsson, Joachim
 Abstract:
 Given a connected geometric graph G, we consider the problem of constructing a tspanner of G having the minimum number of edges. We prove that for every t with 1 1+1/t) edges. This bound almost matches the known upper bound, which states that every connected weighted graph with n vertices contains a tspanner with O(tn1+2/(t+1)) edges. We also prove that the problem of deciding whether a given geometric graph contains a tspanner with at most K edges is NPhard. Previously, this NPhardness result was only known for nongeometric graphs.
 Date Created:
 20060101

 Resource Type:
 Conference Proceeding
 Creator:
 Mirandola, Raffaela, Grassi, Vincenzo, Sabetta, Antonino, and Petriu, Dorina C.
 Abstract:
 The verification of nonfunctional requirements of software models (such as performance, reliability, scalability, security, etc.) requires the transformation of UML models into different analysis models such as Petri nets, queueing networks, formal logic, etc., which represent the system at a higher level of abstraction. The paper proposes a new "abstractionraising" transformation approach for generating analysis models from UML models. In general, such transformations must bridge a large semantic gap between the source and the target model. The proposed approach is illustrated by a transformation from UML to Klaper (Kernel LAnguage for PErformance and Reliability analysis of componentbased systems).
 Date Created:
 20060706

 Resource Type:
 Conference Proceeding
 Creator:
 Dujmović, Vida, Wood, David, and Bose, Prosenjit
 Abstract:
 We prove that for all 0 ≤ t ≤ k and d ≥ 2k, every graph G with treewidth at most k has a 'large' induced subgraph H, where H has treewidth at most t and every vertex in H has degree at most d in G, The order of H depends on t, k, d, and the order of G. With t = k, we obtain large sets of bounded degree vertices. With t = 0, we obtain large independent sets of bounded degree. In both these cases, our bounds on the order of H are tight. For bounded degree independent sets in trees, we characterise the extremal graphs. Finally, we prove that an interval graph with maximum clique size k has a maximum independent set in which every vertex has degree at most 2k.
 Date Created:
 20051201

 Resource Type:
 Conference Proceeding
 Creator:
 Kranakis, Evangelos, Morin, Pat, and Krizanc, Danny
 Abstract:
 We present a tradeoff between the expected time for two identical agents to rendezvous on a synchronous, anonymous, oriented ring and the memory requirements of the agents. In particular, we show that there exists a 2t state agent, which can achieve rendezvous on an n node ring in expected time O( n 2/2 t ∈+∈2 t ) and that any t/2 state agent requires expected time Ω( n 2/2 t ). As a corollary we observe that Θ(loglogn) bits of memory are necessary and sufficient to achieve rendezvous in linear time.
 Date Created:
 20080512

 Resource Type:
 Conference Proceeding
 Creator:
 Markou, Euripides, Kranakis, Evangelos, and Krizanc, Danny
 Abstract:
 We consider the rendezvous problem for identical mobile agents (i.e., running the same deterministic algorithm) with tokens in a synchronous torus with a sense of direction and show that there is a striking computational difference between one and more tokens. More specifically, we show that 1) two agents with a constant number of unmovable tokens, or with one movable token, each cannot rendezvous if they have o(log n) memory, while they can perform rendezvous with detection as long as they have one unmovable token and O(log n) memory; in contrast, 2) when two agents have two movable tokens each then rendezvous (respectively, rendezvous with detection) is possible with constant memory in an arbitrary n × m (respectively, n × n) torus; and finally, 3) two agents with three movable tokens each and constant memory can perform rendezvous with detection in a n × m torus. This is the first publication in the literature that studies tradeoffs between the number of tokens, memory and knowledge the agents need in order to meet in such a network.
 Date Created:
 20060710

 Resource Type:
 Conference Proceeding
 Creator:
 Loke, R. K.S. and Oommen, B. John
 Abstract:
 We consider a problem which can greatly enhance the areas of cursive script recognition and the recognition of printed character sequences. This problem involves recognizing words/strings by processing their noisy subsequences. Let X* be any unknown word from a finite dictionary H. Let U be any arbitrary subsequence of X*. We study the problem of estimating X* by processing Y, a noisy version of U. Y contains substitution, insertion, deletion and generalized transposition errors — the latter occurring when transposed characters are themselves subsequently substituted. We solve the noisy subsequence recognition problem by defining and using the constrained edit distance between X ε H and Y subject to any arbitrary edit constraint involving the number and type of edit operations to be performed. An algorithm to compute this constrained edit distance has been presented. Using these algorithms we present a syntactic Pattern Recognition (PR) scheme which corrects noisy text containing all these types of errors. Experimental results which involve strings of lengths between 40 and 80 with an average of 30.24 deleted characters and an overall average noise of 68.69 % demonstrate the superiority of our system over existing methods.
 Date Created:
 19950101

 Resource Type:
 Conference Proceeding
 Creator:
 Jiang, Lei, Bertossi, Leopoldo, and Rizzolo, Flavio
 Abstract:
 We motivate, formalize and investigate the notions of data quality assessment and data quality query answering as context dependent activities. Contexts for the assessment and usage of a data source at hand are modeled as collections of external databases, that can be materialized or virtual, and mappings within the collections and with the data source at hand. In this way, the context becomes "the complement" of the data source wrt a data integration system. The proposed model allows for natural extensions, like considering data quality predicates, and even more expressive ontologies for data quality assessment.
 Date Created:
 20110101

 Resource Type:
 Conference Proceeding
 Creator:
 Smid, Michiel
 Date Created:
 20091016
 « Previous
 Next »
 1
 2
 3