MACHINE learning techniques can, automatically,

Size: px
Start display at page:

Download "MACHINE learning techniques can, automatically,"

Transcription

1 Proceedings of Internationa Joint Conference on Neura Networks, Daas, Texas, USA, August 4-9, 203 High Leve Data Cassification Based on Network Entropy Fiipe Aves Neto and Liang Zhao Abstract Traditiona data cassification is based ony on physica features of input data. They are caed ow eve cassification. Data cassification by considering not ony physica attributes but aso pattern formation is denominated high eve cassification. In this paper, we propose a new technique that performs high eve cassification by extracting information of networks constructed from the input data. Specificay, we cacuate the network entropies before and after the insertion of a data item to be cassified. Then, we cassify it as beonging to the cass which resuts in the argest increase of the entropy measures. We show that the proposed method can execute cassification tasks according to both simiarity and pattern formation of input data to reach good resuts in the experiments with artificia and rea data sets. In summary, our technique can cacuate how significant a data item is for each cass performing a new way to cassify data. I. INTRODUCTION MACHINE earning techniques can, automaticay, buid modes that are abe to organize the existing knowedge or to mimic specific human behavior by using computing representation of the data retrieved from severa domains []. Machine earning techniques are traditionay divided in two main casses: supervised earning and unsupervised earning [2]. Supervised earning aims to induce concepts from abeed sampes from known casses, that is, the earning process buids an input-output map based on the observation of training data. When the abes consist in discrete vaues, the probem is denominated cassification and, when the vaues are continuous, regression. The constructed map used to predict new data abes is caed cassifier. In unsupervised earning, aso entited custering, the main task is to group the data by some simiarity criteria. Learning process of the custering is ony based on data features, because there is no foreknowedge about the data casses [2], [3]. Artificia Neura Networks, K-Nearest Neighbors (KNN) and Support Vector Machines (SVM) are data cassification techniques broady studied nowadays. However these techniques cassify the data based on the physica features (e.g. simiarity or distance) of the input. This earning process is caed ow eve cassification [4]. Rea data items tend to have an impicit tendency to form patterns in the feature space. We denominate this property as pattern formation, which has a semantic meaning. Because of that, many others researches focus on training and cassifying data considering not ony the physica features but aso the Fiipe Aves Neto and Liang Zhao are with the Department of Computer Sciences, Institute of Mathematics and Computer Science (ICMC), University of São Pauo (USP), Av. Trabahador São-carense, 400, , São Caros, SP, Brazi (emai: fiipeneto@usp.br, zhao@icmc.usp.br). This work is supported by São Pauo State Research Foundation (FAPESP) and the Braziian Nationa Research Counci (CNPq). semantic meaning of the input data, such as co-training technique and Semantic Web, but there is not an expicit and genera scheme to dea with this approach in iterature yet. Data cassification by considering the physica features and the pattern formation is known as high eve cassification [4]. A fundamenta research area in this work is compex network. Formay, a compex network can be represented as a arge scae graph with non trivia connection pattern [5]. Here, the terms network and graph are equivaent. Many natura phenomena can be characterized by a network, ike brain structures, socia interactions and the internet. These systems can be represented by using vertices and edges indicating connections between them. In internet, for exampe, the vertices coud represent routers and the edges its physica connection [6]. Siva and Zhao [4] propose a data cassification technique where a network characterizes the pattern formation of the input data by exporing its topoogica properties. The cassification process, initiay, buids a network for each cass in order to represent each pattern formation, after that, the process verifies the accordance between the pattern formation of the network and the input data instance for cassification. The accordance is verified by the comparison of the network measurements before and after the insertion of the input in the network. Another important topic to this work is random wak which is a fundamenta dynamic process [7]. Given a network and a start point, we seect a neighbor of it at random, and move to this neighbor; then we seect a neighbor of this ast chosen node at random, and move to it, and so on. The sequence of seected points is a random wak [8]. A random wak is a finite Markov chain that is time-reversibe, in other words, every Markov chain can be viewed as random wak on a directed network with weighted edges [8]. Markov chain is the stochastic process defined by a countabe or finite number of states where the conditiona probabiity that the process is in a given state in the current time depends ony on the previous state and not on the sequence of states that preceded it [9]. The matrix that contains the vaues of these probabiities is caed stochastic matrix of the process. In this paper, we propose a high eve data cassification technique based on the extraction of dynamica and topoogica features of the networks constructed from the input data by their entropy. Network entropy is the entropy of a stochastic matrix associated with the adjacency matrix of the network [0] and is reated to the robustness of networks, i.e. their resiience to random changes [6]. Furthermore, network entropy aso characterizes the mutipicity of interna pathways and is negativey correated with the shortest av /3/$ IEEE 2405

2 erage path ength [0]. Consequenty, it captures from oca to goba features of the data, making it ideay suited for high eve cassification. The achieved resuts show that the proposed technique can not ony cassify according to the simiarity, but aso considering the pattern formation. The remainder of the paper is organized as foows. The proposed mode definition is described in section II. In section III, we provide computer simuation resuts to show the particuar characteristics of the proposed technique. Finay, section IV concudes the paper. II. MODEL DESCRIPTION In this section, we present a new cassification technique that anayzes the semantic meaning of the data by using their patterns reationships in networks. Consider a given training set denoted here as X training = {(x,y ),...,(x n,y n )}, where the first component of the ith tupe denotes the attributes of the ith data item. If there are d attributes in each data item, it is said to be a d-dimensiona data set. The second component of the ith tupe characterizes the cass abe associated to the ith data item. The goa in machine earning is to buid a mapping from x to y, this mapping is caed cassifier. The constructed cassifier can be checked by using a test set X test = {x,...,x m }, in which abes are not provided and X training X test =. Here we wi dea ony with discrete abes - y i {,...,L}, i {,...,n} -, so the task is denominated as cassification. From these definitions, the foowing subsections wi describe the cassification technique proposed in this paper. The cassification process is divided in three phases: Network Formation Phase, Entropy Cacuation Phase and Decision Phase. A. Network Formation Phase Initiay, an undirected and weighted network with a singe connected component is buid for each cass. We wi ca G the network that represents the cass {,...,L}. To construct the network G, every data item of the training set, X training, beonging to cass is represented as a vertex. There wi be a connection between two vertices if the Eucidean distance between their feature vectors is ess than a predefined threshod vaue. Moreover, the weight of the connection shoud be inversey proportiona to the distance of the vertices, i.e., the coser the vertices, the stronger must be the connection. So, the weight of the connection between two vertices, i and j, is cacuated by a ij = a ji = ( x i x j )α if if x i x j = 0 0 < x i x j ǫ 0 if x i x j > ǫ Thereby, the constructed network, represented by the adjacency matrix A = (a ij ), preserves the simiarity and the topoogy of the data. The parameter ǫ > 0 determines which vertices shoud be connected and the parameter α 0 indicates how much the distance of the vertices wi be considered, that is, if α 0, the weight of the connections () tends to be binary. It is important to note that the network of each cass needs to have a singe connected component, so the choice of the parameter ǫ must refect this condition. After that, we buid the networks G (i), {,...,L} and i {,...,u} that are based on the data items, beonging to cass, of the training set and the ith data item of the test set aso foowing Equation (), but in the cases when the extra vertex do not have any connection with other vertices, we consider G (i) = G. A these networks wi be used in the next phase of the cassification process. B. Entropy Cacuation Phase In this phase, the entropy of each constructed network is cacuated. Network entropy is the entropy of a stochastic matrix associated with the adjacency matrix A [0]. Then, firsty, we must obtain the stochastic matrix P = (p ij ) associated with these networks. Let λ denote the dominant eigenvaue of A and et (v i ) be the corresponding eading eigenvector. The stochastic matrix is defined by p ij = a ijv j λv i (2) With it at hands, we are abe to cacuate the network entropy by using the dynamica entropy of this stochastic process where p ij describes the transition rates from state i j and its stationary distribution, π = πp. The dynamica entropy, H(P), is defined as H(P) = i π i H i, where H i = j p ij ogp ij (3) Here H i is the Shannon entropy of the distribution [p i,...,p in ] and H is the weighted average over a stationary states [0]. For the next phase, we denominate H(G ) the entropy of the network that represents the cass and H(G (i) ) the entropy of the network G appended the ith item of the test set X test. C. Decision Phase Finay, we wi define the fuzzy cassifier C that decides what cass the data item beong. The hypothesis of the technique is: if the ith data item of the test set X test beongs to cass {,...,L}, the proportiona increase of entropy, (i), is maximum. This idea is expained by the property of networks to dea with random changes, so a data item which does not beong to certain cass wi not affect the respective network. Thus, the item is cassified by its importance in the cass. Mathematicay, the probabiity C(i, ) that the ith data item of the test set X test is cassified as beonging to the cass {,...,L} is given by C(i,) = (i) k L (i) k, where (i) = H(G(i) ) H(G ) (4) 2406

3 The vaues of C(i, ) ies between 0 and and C(i,k) =. k L III. SIMULATION RESULTS In this section, we present three simuations in order to iustrate the efficiency of the proposed high eve cassification method when appied in rea and artificia data sets and compare its resuts with traditiona cassification techniques. Firsty, we use a we known and broady studied database in pattern recognition iterature, the Iris Data Set [], secondariy, other database from UCI, the Seeds Data Set, and, asty, an artificia data set that represents a difficut situation for ow eve cassification Iris Setosa cass Iris Versicoour cass Iris Virginica cass Iris Setosa items from test set (a) A. Simuation by using Iris Data Set Iris is a database which contains 3 casses (Iris Setosa, Iris Versicoour, Iris Virginica) and 50 instances, where each cass refers to a kind of pant. One cass is ineary separabe from the other two, but these are not ineary separabe from each other []. In this simuation, we tested our technique in this database. We chose the parameters ǫ ying between and 2, and α = 0,,2 for Equation () in network formation phase. Figure shows our resuts of this simuation. In this case, we reaize that the smaer the vaues of ǫ, the higher are the accuracy of the cassifier. Moreover, the best resuts are achieved by using ow vaues of α reaching unti 96% of accuracy in 0-fod cross-vaidation. In this way, the proposed technique gets good resuts in this simuation. cassifier accuracy ε α = 0 α = α = 2 Fig.. An anaysis of the cassifier accuracy on different vaues of the parameters ǫ and α appied to the Iris database in 0-fod cross-vaidation. Three different vaues for α were chosen in order to test the cassifier in binary (bue ine with circes) and weighted (green ine with diamonds and red ine with crosses) networks. It is important to notice that vaues ǫ < does not create networks with a singe connected component, thus we do not use these vaues. In order to expicit the cassification detais, Figure 2 shows the proportiona increase of network entropy, (i), due to insertion of data items from test set in the networks of each cass. The parameters are set as: ǫ = and α = 0, and we used 66% of data in training set. Figure 2a shows in detais the ratio between the vaues of network entropy before and after the insertion of the Iris Setosa data item from test set in networks of each cass. In.05 Iris Setosa cass Iris Versicoour cass Iris Virginica cass Iris Versicoour items from test set Iris Setosa cass Iris Versicoour cass Iris Virginica cass (b) Iris Virginica items from test set Fig. 2. Comparison of the proportiona increase of network entropy for each cass by inserting the data items from test set. Networks that represent the casses are buit by using 66% of data and parameters ǫ = and α = 0. Each figure shows the resuts obtained by inserting data items from test set, where Figure 2a expose the Iris Setosa items, Figure 2b the Iris Versicoour items and Figure 2c the Iris Virginica items. Bue ines indicate the vaues (i) IrisSetosa, i. Green and red ines indicate, respectivey, (i) IrisV ersicoour and (i) IrisV irginica, i. this figure, we observe that these data items increased the entropy ony in the network of the same cass. It happens because the Iris Setosa cass is ineary separabe of the other two casses. Simiary, the increase of the network entropy due to Iris Versicoour data items from test set is showed in Figure 2b. Finay, Figure 2c present the same resuts by inserting the Iris Virginica data items from test set. Emphasized in this figure, we can see the data items that are miscassified as beonging to Iris Versicoour cass. (c) B. Simuation by using Seeds Data Set Seeds database contains 3 casses (Kama, Rosa, Canadian), 20 instances and 7 attributes for each data item. The casses refer to three different varieties of wheat []. In this simuation, we provide an anaysis simiar to previous 2407

4 subsection for this database. Here, we chose the parameters ǫ ying between 2.5 and 3.5, and α = 0,,2 for Equation () in network formation phase. Figure 3 shows our resuts of this simuation. In this case, the best resut was achieved by using α = 0 and ǫ = 3. reaching 9% of accuracy in 0-fod cross-vaidation. Like in the ast simuation, some traditiona cassification techniques reach simiar resuts to ours Kama cass Rosa cass Canadian cass (a) cassifier accuracy α = 0 α = α = ε Kama cass Rosa cass Canadian cass Fig. 3. An anaysis of the cassifier accuracy on different vaues of the parameters ǫ and α appied to the Seeds database in 0-fod cross-vaidation. Three different vaues for α were chosen in order to test the cassifier in binary (bue ine with circes) and weighted (green ine with diamonds and red ine with crosses) networks (b) It is important to notice that vaues ǫ < 2.5 can not create networks with a singe connected component. In order to expicit the cassification detais, Figure 4 shows the proportiona increase of network entropy, (i), due to insertion of data items from test set in the networks of each cass. The parameters are set as: ǫ = 3. and α = 0, and we used 66% of data in training set. Figure 4a shows in detais the ratio between the vaues of network entropy before and after the insertion of the Kama data item from test set in networks of each cass. Simiary, the increase of the network entropy due to Rosa data items from test set is showed in Figure 4b. The data items that were incorrecty cassified are highighted in these both figure. Finay, Figure 4c present the same resuts by inserting the Canadian data items from test set. In this figure, we observe that a Canadian data items were correcty cassified. C. Simuation by using an Artificia Data Set In this simuation, we tested our cassification method in a situation where traditiona techniques can not extract enough information to correcty cassify the data. The artificia data set is described by Figure 5. The database consists in two casses forming two grids where approximatey 4% of them are intersected. Despite the difficut of this probem, our cassification method obtained better resuts than other techniques. Figure 6 demonstrate how we each one of the techniques cassify the data with different sizes of the training set. The traditiona cassification techniques used here are Mutiayer Perceptron (MLP) and Support Vector Machine (SVM) Kama cass Rosa cass Canadian cass Fig. 4. Comparison of the proportiona increase of network entropy for each cass by inserting the data items from test set. Networks that represent the casses are buit by using 66% of data and parameters ǫ = 3. and α = 0. Each figure shows the resuts obtained by inserting data items from test set, where Figure 4a expose the Kama items, Figure 4b the Rosa items and Figure 4c the Canadian items. Bue ines indicate the vaues (i) Kama, i. Green and red ines indicate, respectivey, (i) Rosa and (i) Canadian, i. The MLP cassifier obtained accuracy between 96.3% and 97.5% whie SVM obtained cassification rate between 90.4% and 96.3%. It is worth mentioning that these vaues of accuracy are not good because more than 92% of data in this database coud be considered easy to be cassified. Our technique, in contrast, coud cassify correcty more than 98% of data in test set by using ony 40% of a data in training set and 00% of accuracy with 90% of data in training set. Besides, the bigger are the training set size, the better are the resuts of the appication of our technique whist it does not happen in the other techniques appications. (c) IV. CONCLUSIONS Here, we propose a high eve data cassification technique based on network entropy. This method cassifies the data 2408

5 Fig. 5. Representation of the artificia database constructed to evauate our technique. The database has two bi-dimensiona casses that intersect themseves. The casses has a simiar pattern formation, in fact, both casses are grids with distance between two adjacent data item, but the red cass data items are shifted in reation to bue cass data items. items by their significance in the network that represent the casses. As a resut, the technique can cassify according the physica and semantic features of the input data. The choice of the parameters, ǫ and α, in the network formation phase is crucia to the cassifier achieves good accuracy. An interesting fact in this work is that the bigger the vaues of the parameters, the bigger is the amount of information you obtain from data, but not aways a huge amount of information obtain better resuts. For exampe, the seection of the parameters in the simuation by using Iris database was α = 0 and ǫ = to focus at distance of the data, i.e., the resuting network is binary with few connections, and this approach achieved better resuts than other. Furthermore, we expose a situation (subsection III-C) which provides evidences that increasing the size of training set in traditiona data cassification techniques can not improve accuracy or induce errors when the simiarity is not enough to cassify the data since these techniques consider ony oca aspects. This probem can be overcome when the proposed technique is used. Nevertheess, there are some imitations in this technique. Firsty, the agorithm is not scaabe to arge data sets because the computationa costs of constructing networks and cacuating entropies are very expensive. Secondy, imbaanced data set may have unexpected effects on this method. Lasty, the performance of the technique is very sensitive to the parameters. Concuding, this work is on track to improve current resuts in rea data cassification appications presenting innovation for machine earning theory. cassifier accuracy 4 2 Network Entropy SVM MLP training set size Fig. 6. A detaied study of our technique performance compared to other traditiona techniques appied in the artificia database described by Figure 5 varying the size of the training set used. The bue ine with circes indicates the resuts of our technique appication with parameters ǫ = and α =. The red ine with crosses shows the performance reached by a Mutiayer Perceptron cassifier with 4 neuron in output ayer and earning rate 0.3. The green ine with diamonds depicts the resuts achieved by a Support Vector Machine cassifier with a radia basis function as kerne. REFERENCES [] C.M. Bishop, Pattern recognition and machine earning. Springer, [2] T.M. Mitche, Machine earning. McGraw-Hi Series in Computer Science. McGraw-Hi, 997. [3] S.J. Russe and P. Norvig, Artificia Inteigence: A Modern Approach. pp , Prentice Ha, 3 edition, 200. [4] T.C. Siva and L. Zhao, Network-Based High Leve Data Cassification. IEEE Transactions on Neura Networks and Learning Systems, v. 23, pp , 202. [5] S. Boccaetti, V. Latora, Y. Moreno, M. Chavez and D.U. Hwang, Compex networks: Structure and dynamics. Physics Reports, vo. 424, pp , [6] L.F. Costa, F.A. Rodrigues, G. Travieso and P.R. Vias, Characterization of compex networks: A survey of measurements. Advances in Physics, vo. 56, pp , [7] J.D. Noh and H. Rieger, Random Waks on Compex Networks. Physica Review Letters, vo. 92, pp. 870, [8] L. Lovász, Random Waks on Graphs: a Survey. Yae University, 994. [9] S.M. Ross, Introduction to probabiity modes. Academic Press, 0 edition, [0] L. Demetrius and T. Manke, Robustness and network evoution - an entropic principe. Physica A, vo. 346, pp , [] A. Frank and A. Asuncion. UCI Machine Learning Repository [ Irvine, CA: University of Caifornia, Schoo of Information and Computer Science

Nearest Neighbor Learning

Nearest Neighbor Learning Nearest Neighbor Learning Cassify based on oca simiarity Ranges from simpe nearest neighbor to case-based and anaogica reasoning Use oca information near the current query instance to decide the cassification

More information

Language Identification for Texts Written in Transliteration

Language Identification for Texts Written in Transliteration Language Identification for Texts Written in Transiteration Andrey Chepovskiy, Sergey Gusev, Margarita Kurbatova Higher Schoo of Economics, Data Anaysis and Artificia Inteigence Department, Pokrovskiy

More information

A New Supervised Clustering Algorithm Based on Min-Max Modular Network with Gaussian-Zero-Crossing Functions

A New Supervised Clustering Algorithm Based on Min-Max Modular Network with Gaussian-Zero-Crossing Functions 2006 Internationa Joint Conference on Neura Networks Sheraton Vancouver Wa Centre Hote, Vancouver, BC, Canada Juy 16-21, 2006 A New Supervised Custering Agorithm Based on Min-Max Moduar Network with Gaussian-Zero-Crossing

More information

Mobile App Recommendation: Maximize the Total App Downloads

Mobile App Recommendation: Maximize the Total App Downloads Mobie App Recommendation: Maximize the Tota App Downoads Zhuohua Chen Schoo of Economics and Management Tsinghua University chenzhh3.12@sem.tsinghua.edu.cn Yinghui (Catherine) Yang Graduate Schoo of Management

More information

Image Segmentation Using Semi-Supervised k-means

Image Segmentation Using Semi-Supervised k-means I J C T A, 9(34) 2016, pp. 595-601 Internationa Science Press Image Segmentation Using Semi-Supervised k-means Reza Monsefi * and Saeed Zahedi * ABSTRACT Extracting the region of interest is a very chaenging

More information

Research of Classification based on Deep Neural Network

Research of  Classification based on Deep Neural Network 2018 Internationa Conference on Sensor Network and Computer Engineering (ICSNCE 2018) Research of Emai Cassification based on Deep Neura Network Wang Yawen Schoo of Computer Science and Engineering Xi

More information

Automatic Hidden Web Database Classification

Automatic Hidden Web Database Classification Automatic idden Web atabase Cassification Zhiguo Gong, Jingbai Zhang, and Qian Liu Facuty of Science and Technoogy niversity of Macau Macao, PRC {fstzgg,ma46597,ma46620}@umac.mo Abstract. In this paper,

More information

Sensitivity Analysis of Hopfield Neural Network in Classifying Natural RGB Color Space

Sensitivity Analysis of Hopfield Neural Network in Classifying Natural RGB Color Space Sensitivity Anaysis of Hopfied Neura Network in Cassifying Natura RGB Coor Space Department of Computer Science University of Sharjah UAE rsammouda@sharjah.ac.ae Abstract: - This paper presents a study

More information

ACTIVE LEARNING ON WEIGHTED GRAPHS USING ADAPTIVE AND NON-ADAPTIVE APPROACHES. Eyal En Gad, Akshay Gadde, A. Salman Avestimehr and Antonio Ortega

ACTIVE LEARNING ON WEIGHTED GRAPHS USING ADAPTIVE AND NON-ADAPTIVE APPROACHES. Eyal En Gad, Akshay Gadde, A. Salman Avestimehr and Antonio Ortega ACTIVE LEARNING ON WEIGHTED GRAPHS USING ADAPTIVE AND NON-ADAPTIVE APPROACHES Eya En Gad, Akshay Gadde, A. Saman Avestimehr and Antonio Ortega Department of Eectrica Engineering University of Southern

More information

Interpreting Individual Classifications of Hierarchical Networks

Interpreting Individual Classifications of Hierarchical Networks Interpreting Individua Cassifications of Hierarchica Networks Wi Landecker, Michae D. Thomure, Luís M. A. Bettencourt, Meanie Mitche, Garrett T. Kenyon, and Steven P. Brumby Department of Computer Science

More information

Distance Weighted Discrimination and Second Order Cone Programming

Distance Weighted Discrimination and Second Order Cone Programming Distance Weighted Discrimination and Second Order Cone Programming Hanwen Huang, Xiaosun Lu, Yufeng Liu, J. S. Marron, Perry Haaand Apri 3, 2012 1 Introduction This vignette demonstrates the utiity and

More information

A Design Method for Optimal Truss Structures with Certain Redundancy Based on Combinatorial Rigidity Theory

A Design Method for Optimal Truss Structures with Certain Redundancy Based on Combinatorial Rigidity Theory 0 th Word Congress on Structura and Mutidiscipinary Optimization May 9 -, 03, Orando, Forida, USA A Design Method for Optima Truss Structures with Certain Redundancy Based on Combinatoria Rigidity Theory

More information

Automatic Grouping for Social Networks CS229 Project Report

Automatic Grouping for Social Networks CS229 Project Report Automatic Grouping for Socia Networks CS229 Project Report Xiaoying Tian Ya Le Yangru Fang Abstract Socia networking sites aow users to manuay categorize their friends, but it is aborious to construct

More information

Massively Parallel Part of Speech Tagging Using. Min-Max Modular Neural Networks.

Massively Parallel Part of Speech Tagging Using. Min-Max Modular Neural Networks. assivey Parae Part of Speech Tagging Using in-ax oduar Neura Networks Bao-Liang Lu y, Qing a z, ichinori Ichikawa y, & Hitoshi Isahara z y Lab. for Brain-Operative Device, Brain Science Institute, RIEN

More information

Optimization and Application of Support Vector Machine Based on SVM Algorithm Parameters

Optimization and Application of Support Vector Machine Based on SVM Algorithm Parameters Optimization and Appication of Support Vector Machine Based on SVM Agorithm Parameters YAN Hui-feng 1, WANG Wei-feng 1, LIU Jie 2 1 ChongQing University of Posts and Teecom 400065, China 2 Schoo Of Civi

More information

Improvement of Nearest-Neighbor Classifiers via Support Vector Machines

Improvement of Nearest-Neighbor Classifiers via Support Vector Machines From: FLAIRS-01 Proceedings. Copyright 2001, AAAI (www.aaai.org). A rights reserved. Improvement of Nearest-Neighbor Cassifiers via Support Vector Machines Marc Sebban and Richard Nock TRIVIA-Department

More information

Interpreting Individual Classifications of Hierarchical Networks

Interpreting Individual Classifications of Hierarchical Networks Portand State University PDXSchoar Computer Science Facuty Pubications and Presentations Computer Science 2013 Interpreting Individua Cassifications of Hierarchica Networks Wi Landecker Portand State University,

More information

FREE-FORM ANISOTROPY: A NEW METHOD FOR CRACK DETECTION ON PAVEMENT SURFACE IMAGES

FREE-FORM ANISOTROPY: A NEW METHOD FOR CRACK DETECTION ON PAVEMENT SURFACE IMAGES FREE-FORM ANISOTROPY: A NEW METHOD FOR CRACK DETECTION ON PAVEMENT SURFACE IMAGES Tien Sy Nguyen, Stéphane Begot, Forent Ducuty, Manue Avia To cite this version: Tien Sy Nguyen, Stéphane Begot, Forent

More information

DETERMINING INTUITIONISTIC FUZZY DEGREE OF OVERLAPPING OF COMPUTATION AND COMMUNICATION IN PARALLEL APPLICATIONS USING GENERALIZED NETS

DETERMINING INTUITIONISTIC FUZZY DEGREE OF OVERLAPPING OF COMPUTATION AND COMMUNICATION IN PARALLEL APPLICATIONS USING GENERALIZED NETS DETERMINING INTUITIONISTIC FUZZY DEGREE OF OVERLAPPING OF COMPUTATION AND COMMUNICATION IN PARALLEL APPLICATIONS USING GENERALIZED NETS Pave Tchesmedjiev, Peter Vassiev Centre for Biomedica Engineering,

More information

Lecture Notes for Chapter 4 Part III. Introduction to Data Mining

Lecture Notes for Chapter 4 Part III. Introduction to Data Mining Data Mining Cassification: Basic Concepts, Decision Trees, and Mode Evauation Lecture Notes for Chapter 4 Part III Introduction to Data Mining by Tan, Steinbach, Kumar Adapted by Qiang Yang (2010) Tan,Steinbach,

More information

TSR: Topology Reduction from Tree to Star Data Grids

TSR: Topology Reduction from Tree to Star Data Grids 03 Seventh Internationa Conference on Innovative Mobie and Internet Services in biquitous Computing TSR: Topoogy Reduction from Tree to Star Data Grids Ming-Chang Lee #, Fang-Yie Leu *, Ying-ping Chen

More information

A Petrel Plugin for Surface Modeling

A Petrel Plugin for Surface Modeling A Petre Pugin for Surface Modeing R. M. Hassanpour, S. H. Derakhshan and C. V. Deutsch Structure and thickness uncertainty are important components of any uncertainty study. The exact ocations of the geoogica

More information

Hiding secrete data in compressed images using histogram analysis

Hiding secrete data in compressed images using histogram analysis University of Woongong Research Onine University of Woongong in Dubai - Papers University of Woongong in Dubai 2 iding secrete data in compressed images using histogram anaysis Farhad Keissarian University

More information

Alpha labelings of straight simple polyominal caterpillars

Alpha labelings of straight simple polyominal caterpillars Apha abeings of straight simpe poyomina caterpiars Daibor Froncek, O Nei Kingston, Kye Vezina Department of Mathematics and Statistics University of Minnesota Duuth University Drive Duuth, MN 82-3, U.S.A.

More information

Backing-up Fuzzy Control of a Truck-trailer Equipped with a Kingpin Sliding Mechanism

Backing-up Fuzzy Control of a Truck-trailer Equipped with a Kingpin Sliding Mechanism Backing-up Fuzzy Contro of a Truck-traier Equipped with a Kingpin Siding Mechanism G. Siamantas and S. Manesis Eectrica & Computer Engineering Dept., University of Patras, Patras, Greece gsiama@upatras.gr;stam.manesis@ece.upatras.gr

More information

Neural Network Enhancement of the Los Alamos Force Deployment Estimator

Neural Network Enhancement of the Los Alamos Force Deployment Estimator Missouri University of Science and Technoogy Schoars' Mine Eectrica and Computer Engineering Facuty Research & Creative Works Eectrica and Computer Engineering 1-1-1994 Neura Network Enhancement of the

More information

A HYBRID FEATURE SELECTION METHOD BASED ON FISHER SCORE AND GENETIC ALGORITHM

A HYBRID FEATURE SELECTION METHOD BASED ON FISHER SCORE AND GENETIC ALGORITHM Journa of Mathematica Sciences: Advances and Appications Voume 37, 2016, Pages 51-78 Avaiabe at http://scientificadvances.co.in DOI: http://dx.doi.org/10.18642/jmsaa_7100121627 A HYBRID FEATURE SELECTION

More information

A NEW APPROACH FOR BLOCK BASED STEGANALYSIS USING A MULTI-CLASSIFIER

A NEW APPROACH FOR BLOCK BASED STEGANALYSIS USING A MULTI-CLASSIFIER Internationa Journa on Technica and Physica Probems of Engineering (IJTPE) Pubished by Internationa Organization of IOTPE ISSN 077-358 IJTPE Journa www.iotpe.com ijtpe@iotpe.com September 014 Issue 0 Voume

More information

understood as processors that match AST patterns of the source language and translate them into patterns in the target language.

understood as processors that match AST patterns of the source language and translate them into patterns in the target language. A Basic Compier At a fundamenta eve compiers can be understood as processors that match AST patterns of the source anguage and transate them into patterns in the target anguage. Here we wi ook at a basic

More information

AUTOMATIC gender classification based on facial images

AUTOMATIC gender classification based on facial images SUBMITTED TO IEEE TRANSACTIONS ON NEURAL NETWORKS 1 Gender Cassification Using a Min-Max Moduar Support Vector Machine with Incorporating Prior Knowedge Hui-Cheng Lian and Bao-Liang Lu, Senior Member,

More information

Outline. Introduce yourself!! What is Machine Learning? What is CAP-5610 about? Class information and logistics

Outline. Introduce yourself!! What is Machine Learning? What is CAP-5610 about? Class information and logistics Outine Introduce yoursef!! What is Machine Learning? What is CAP-5610 about? Cass information and ogistics Lecture Notes for E Apaydın 2010 Introduction to Machine Learning 2e The MIT Press (V1.0) About

More information

Fast Methods for Kernel-based Text Analysis

Fast Methods for Kernel-based Text Analysis Proceedings of the 41st Annua Meeting of the Association for Computationa Linguistics, Juy 2003, pp. 24-31. Fast Methods for Kerne-based Text Anaysis Taku Kudo and Yuji Matsumoto Graduate Schoo of Information

More information

Binarized support vector machines

Binarized support vector machines Universidad Caros III de Madrid Repositorio instituciona e-archivo Departamento de Estadística http://e-archivo.uc3m.es DES - Working Papers. Statistics and Econometrics. WS 2007-11 Binarized support vector

More information

Fuzzy Equivalence Relation Based Clustering and Its Use to Restructuring Websites Hyperlinks and Web Pages

Fuzzy Equivalence Relation Based Clustering and Its Use to Restructuring Websites Hyperlinks and Web Pages Fuzzy Equivaence Reation Based Custering and Its Use to Restructuring Websites Hyperinks and Web Pages Dimitris K. Kardaras,*, Xenia J. Mamakou, and Bi Karakostas 2 Business Informatics Laboratory, Dept.

More information

The Classification of Stored Grain Pests based on Convolutional Neural Network

The Classification of Stored Grain Pests based on Convolutional Neural Network 2017 2nd Internationa Conference on Mechatronics and Information Technoogy (ICMIT 2017) The Cassification of Stored Grain Pests based on Convoutiona Neura Network Dexian Zhang1, Wenun Zhao*, 1 1 Schoo

More information

A Novel Method for Early Software Quality Prediction Based on Support Vector Machine

A Novel Method for Early Software Quality Prediction Based on Support Vector Machine A Nove Method for Eary Software Quaity Prediction Based on Support Vector Machine Fei Xing 1,PingGuo 1;2, and Michae R. Lyu 2 1 Department of Computer Science Beijing Norma University, Beijing, 1875, China

More information

A Novel Linear-Polynomial Kernel to Construct Support Vector Machines for Speech Recognition

A Novel Linear-Polynomial Kernel to Construct Support Vector Machines for Speech Recognition Journa of Computer Science 7 (7): 99-996, 20 ISSN 549-3636 20 Science Pubications A Nove Linear-Poynomia Kerne to Construct Support Vector Machines for Speech Recognition Bawant A. Sonkambe and 2 D.D.

More information

Community-Aware Opportunistic Routing in Mobile Social Networks

Community-Aware Opportunistic Routing in Mobile Social Networks IEEE TRANSACTIONS ON COMPUTERS VOL:PP NO:99 YEAR 213 Community-Aware Opportunistic Routing in Mobie Socia Networks Mingjun Xiao, Member, IEEE Jie Wu, Feow, IEEE, and Liusheng Huang, Member, IEEE Abstract

More information

QoS-Aware Data Transmission and Wireless Energy Transfer: Performance Modeling and Optimization

QoS-Aware Data Transmission and Wireless Energy Transfer: Performance Modeling and Optimization QoS-Aware Data Transmission and Wireess Energy Transfer: Performance Modeing and Optimization Dusit Niyato, Ping Wang, Yeow Wai Leong, and Tan Hwee Pink Schoo of Computer Engineering, Nanyang Technoogica

More information

Testing Whether a Set of Code Words Satisfies a Given Set of Constraints *

Testing Whether a Set of Code Words Satisfies a Given Set of Constraints * JOURNAL OF INFORMATION SCIENCE AND ENGINEERING 6, 333-346 (010) Testing Whether a Set of Code Words Satisfies a Given Set of Constraints * HSIN-WEN WEI, WAN-CHEN LU, PEI-CHI HUANG, WEI-KUAN SHIH AND MING-YANG

More information

Fastest-Path Computation

Fastest-Path Computation Fastest-Path Computation DONGHUI ZHANG Coege of Computer & Information Science Northeastern University Synonyms fastest route; driving direction Definition In the United states, ony 9.% of the househods

More information

Quaternion Support Vector Classifier

Quaternion Support Vector Classifier Quaternion Support Vector Cassifier G. López-Gonzáez, Nancy Arana-Danie, and Eduardo Bayro-Corrochano CINVESTAV - Unidad Guadaajara, Av. de Bosque 1145, Coonia e Bajo, Zapopan, Jaisco, México {geopez,edb}@gd.cinvestav.mx

More information

WATERMARKING GIS DATA FOR DIGITAL MAP COPYRIGHT PROTECTION

WATERMARKING GIS DATA FOR DIGITAL MAP COPYRIGHT PROTECTION WATERMARKING GIS DATA FOR DIGITAL MAP COPYRIGHT PROTECTION Shen Tao Chinese Academy of Surveying and Mapping, Beijing 100039, China shentao@casm.ac.cn Xu Dehe Institute of resources and environment, North

More information

Real-Time Feature Descriptor Matching via a Multi-Resolution Exhaustive Search Method

Real-Time Feature Descriptor Matching via a Multi-Resolution Exhaustive Search Method 297 Rea-Time Feature escriptor Matching via a Muti-Resoution Ehaustive Search Method Chi-Yi Tsai, An-Hung Tsao, and Chuan-Wei Wang epartment of Eectrica Engineering, Tamang University, New Taipei City,

More information

Space-Time Trade-offs.

Space-Time Trade-offs. Space-Time Trade-offs. Chethan Kamath 03.07.2017 1 Motivation An important question in the study of computation is how to best use the registers in a CPU. In most cases, the amount of registers avaiabe

More information

Response Surface Model Updating for Nonlinear Structures

Response Surface Model Updating for Nonlinear Structures Response Surface Mode Updating for Noninear Structures Gonaz Shahidi a, Shamim Pakzad b a PhD Student, Department of Civi and Environmenta Engineering, Lehigh University, ATLSS Engineering Research Center,

More information

Solutions to the Final Exam

Solutions to the Final Exam CS/Math 24: Intro to Discrete Math 5//2 Instructor: Dieter van Mekebeek Soutions to the Fina Exam Probem Let D be the set of a peope. From the definition of R we see that (x, y) R if and ony if x is a

More information

Topology-aware Key Management Schemes for Wireless Multicast

Topology-aware Key Management Schemes for Wireless Multicast Topoogy-aware Key Management Schemes for Wireess Muticast Yan Sun, Wade Trappe,andK.J.RayLiu Department of Eectrica and Computer Engineering, University of Maryand, Coege Park Emai: ysun, kjriu@gue.umd.edu

More information

Quality Assessment using Tone Mapping Algorithm

Quality Assessment using Tone Mapping Algorithm Quaity Assessment using Tone Mapping Agorithm Nandiki.pushpa atha, Kuriti.Rajendra Prasad Research Schoar, Assistant Professor, Vignan s institute of engineering for women, Visakhapatnam, Andhra Pradesh,

More information

Collinearity and Coplanarity Constraints for Structure from Motion

Collinearity and Coplanarity Constraints for Structure from Motion Coinearity and Copanarity Constraints for Structure from Motion Gang Liu 1, Reinhard Kette 2, and Bodo Rosenhahn 3 1 Institute of Information Sciences and Technoogy, Massey University, New Zeaand, Department

More information

Intro to Programming & C Why Program? 1.2 Computer Systems: Hardware and Software. Why Learn to Program?

Intro to Programming & C Why Program? 1.2 Computer Systems: Hardware and Software. Why Learn to Program? Intro to Programming & C++ Unit 1 Sections 1.1-3 and 2.1-10, 2.12-13, 2.15-17 CS 1428 Spring 2018 Ji Seaman 1.1 Why Program? Computer programmabe machine designed to foow instructions Program a set of

More information

Sparse Representation based Face Recognition with Limited Labeled Samples

Sparse Representation based Face Recognition with Limited Labeled Samples Sparse Representation based Face Recognition with Limited Labeed Sampes Vijay Kumar, Anoop Namboodiri, C.V. Jawahar Center for Visua Information Technoogy, IIIT Hyderabad, India Abstract Sparse representations

More information

A Two-Step Approach to Hallucinating Faces: Global Parametric Model and Local Nonparametric Model

A Two-Step Approach to Hallucinating Faces: Global Parametric Model and Local Nonparametric Model A Two-Step Approach to aucinating Faces: Goba Parametric Mode and Loca Nonparametric Mode Ce Liu eung-yeung Shum Chang-Shui Zhang State Key Lab of nteigent Technoogy and Systems, Dept. of Automation, Tsinghua

More information

University of Illinois at Urbana-Champaign, Urbana, IL 61801, /11/$ IEEE 162

University of Illinois at Urbana-Champaign, Urbana, IL 61801, /11/$ IEEE 162 oward Efficient Spatia Variation Decomposition via Sparse Regression Wangyang Zhang, Karthik Baakrishnan, Xin Li, Duane Boning and Rob Rutenbar 3 Carnegie Meon University, Pittsburgh, PA 53, wangyan@ece.cmu.edu,

More information

Semi-Supervised Learning with Sparse Distributed Representations

Semi-Supervised Learning with Sparse Distributed Representations Semi-Supervised Learning with Sparse Distributed Representations David Zieger dzieger@stanford.edu CS 229 Fina Project 1 Introduction For many machine earning appications, abeed data may be very difficut

More information

A probabilistic fuzzy method for emitter identification based on genetic algorithm

A probabilistic fuzzy method for emitter identification based on genetic algorithm A probabitic fuzzy method for emitter identification based on genetic agorithm Xia Chen, Weidong Hu, Hongwen Yang, Min Tang ATR Key Lab, Coege of Eectronic Science and Engineering Nationa University of

More information

Forgot to compute the new centroids (-1); error in centroid computations (-1); incorrect clustering results (-2 points); more than 2 errors: 0 points.

Forgot to compute the new centroids (-1); error in centroid computations (-1); incorrect clustering results (-2 points); more than 2 errors: 0 points. Probem 1 a. K means is ony capabe of discovering shapes that are convex poygons [1] Cannot discover X shape because X is not convex. [1] DBSCAN can discover X shape. [1] b. K-means is prototype based and

More information

JOINT IMAGE REGISTRATION AND EXAMPLE-BASED SUPER-RESOLUTION ALGORITHM

JOINT IMAGE REGISTRATION AND EXAMPLE-BASED SUPER-RESOLUTION ALGORITHM JOINT IMAGE REGISTRATION AND AMPLE-BASED SUPER-RESOLUTION ALGORITHM Hyo-Song Kim, Jeyong Shin, and Rae-Hong Park Department of Eectronic Engineering, Schoo of Engineering, Sogang University 35 Baekbeom-ro,

More information

Comparative Analysis of Relevance for SVM-Based Interactive Document Retrieval

Comparative Analysis of Relevance for SVM-Based Interactive Document Retrieval Comparative Anaysis for SVM-Based Interactive Document Retrieva Paper: Comparative Anaysis of Reevance for SVM-Based Interactive Document Retrieva Hiroshi Murata, Takashi Onoda, and Seiji Yamada Centra

More information

Understanding the Mixing Patterns of Social Networks: The Impact of Cores, Link Directions, and Dynamics

Understanding the Mixing Patterns of Social Networks: The Impact of Cores, Link Directions, and Dynamics Understanding the Mixing Patterns of Socia Networks: The Impact of Cores, Link Directions, and Dynamics [Last revised on May 22, 2011] Abedeaziz Mohaisen Huy Tran Nichoas Hopper Yongdae Kim University

More information

Model-driven Collaboration and Information Integration for Enhancing Video Semantic Concept Detection

Model-driven Collaboration and Information Integration for Enhancing Video Semantic Concept Detection Mode-driven Coaboration and Information Integration for Enhancing Video Semantic Concept Detection Tao Meng, Mei-Ling Shyu Department of Eectrica and Computer Engineering University of Miami Cora Gabes,

More information

Modification of Support Vector Machine for Microarray Data Analysis

Modification of Support Vector Machine for Microarray Data Analysis Goba Journa of Computer Science and Technoogy Hardware & Computation Voume 13 Issue 1 Version 1.0 Year 2013 Type: Doube Bind Peer Reviewed Internationa Research Journa Pubisher: Goba Journas Inc. (USA)

More information

A Method for Group Extraction in Complex Social Networks

A Method for Group Extraction in Complex Social Networks This is not the fina version of this paper. You can find the fina version on the pubisher web page. A Method for Group Extraction in Compex Socia Networks Piotr Bródka 1, Katarzyna Musia 2, Przemysław

More information

Efficient Histogram-based Indexing for Video Copy Detection

Efficient Histogram-based Indexing for Video Copy Detection Efficient Histogram-based Indexing for Video Copy Detection Chih-Yi Chiu, Jenq-Haur Wang*, and Hung-Chi Chang Institute of Information Science, Academia Sinica, Taiwan *Department of Computer Science and

More information

Transformation Invariance in Pattern Recognition: Tangent Distance and Propagation

Transformation Invariance in Pattern Recognition: Tangent Distance and Propagation Transformation Invariance in Pattern Recognition: Tangent Distance and Propagation Patrice Y. Simard, 1 Yann A. Le Cun, 2 John S. Denker, 2 Bernard Victorri 3 1 Microsoft Research, 1 Microsoft Way, Redmond,

More information

5940 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 13, NO. 11, NOVEMBER 2014

5940 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 13, NO. 11, NOVEMBER 2014 5940 IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, VOL. 13, NO. 11, NOVEMBER 014 Topoogy-Transparent Scheduing in Mobie Ad Hoc Networks With Mutipe Packet Reception Capabiity Yiming Liu, Member, IEEE,

More information

FACE RECOGNITION WITH HARMONIC DE-LIGHTING. s: {lyqing, sgshan, wgao}jdl.ac.cn

FACE RECOGNITION WITH HARMONIC DE-LIGHTING.  s: {lyqing, sgshan, wgao}jdl.ac.cn FACE RECOGNITION WITH HARMONIC DE-LIGHTING Laiyun Qing 1,, Shiguang Shan, Wen Gao 1, 1 Graduate Schoo, CAS, Beijing, China, 100080 ICT-ISVISION Joint R&D Laboratory for Face Recognition, CAS, Beijing,

More information

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc.

An Indian Journal FULL PAPER ABSTRACT KEYWORDS. Trade Science Inc. [Type text] [Type text] [Type text] ISSN : 0974-7435 Voume 10 Issue 16 BioTechnoogy 014 An Indian Journa FULL PAPER BTAIJ, 10(16), 014 [999-9307] Study on prediction of type- fuzzy ogic power system based

More information

T he increasing importance of Research and Development (R&D) activities has spurred the formation of

T he increasing importance of Research and Development (R&D) activities has spurred the formation of OPEN SUBJECT AREAS: STATISTICS COMPLEX NETWORKS Received 8 Apri 2014 Accepted 18 June 2014 Pubished 14 Juy 2014 Correspondence and requests for materias shoud be addressed to M.V.T. (mtomaseo@ ethz.ch)

More information

Constellation Models for Recognition of Generic Objects

Constellation Models for Recognition of Generic Objects 1 Consteation Modes for Recognition of Generic Objects Wei Zhang Schoo of Eectrica Engineering and Computer Science Oregon State University zhangwe@eecs.oregonstate.edu Abstract Recognition of generic

More information

Development of a hybrid K-means-expectation maximization clustering algorithm

Development of a hybrid K-means-expectation maximization clustering algorithm Journa of Computations & Modeing, vo., no.4, 0, -3 ISSN: 79-765 (print, 79-8850 (onine Scienpress Ltd, 0 Deveopment of a hybrid K-means-expectation maximization custering agorithm Adigun Abimboa Adebisi,

More information

Delay Budget Partitioning to Maximize Network Resource Usage Efficiency

Delay Budget Partitioning to Maximize Network Resource Usage Efficiency Deay Budget Partitioning to Maximize Network Resource Usage Efficiency Kartik Gopaan Tzi-cker Chiueh Yow-Jian Lin Forida State University Stony Brook University Tecordia Technoogies kartik@cs.fsu.edu chiueh@cs.sunysb.edu

More information

An Introduction to Design Patterns

An Introduction to Design Patterns An Introduction to Design Patterns 1 Definitions A pattern is a recurring soution to a standard probem, in a context. Christopher Aexander, a professor of architecture Why woud what a prof of architecture

More information

On-Chip CNN Accelerator for Image Super-Resolution

On-Chip CNN Accelerator for Image Super-Resolution On-Chip CNN Acceerator for Image Super-Resoution Jung-Woo Chang and Suk-Ju Kang Dept. of Eectronic Engineering, Sogang University, Seou, South Korea {zwzang91, sjkang}@sogang.ac.kr ABSTRACT To impement

More information

IJOART. Copyright 2013 SciResPub. Shri Venkateshwara University, Institute of Computer & Information Science. India. Uttar Pradesh, India

IJOART. Copyright 2013 SciResPub. Shri Venkateshwara University, Institute of Computer & Information Science. India. Uttar Pradesh, India Internationa Journa of Advancements in Research & Technoogy, Voume, Issue 6, June-04 ISSN 78-7763 89 Pattern assification for andwritten Marathi haracters using Gradient Descent of Distributed error with

More information

A Robust Sign Language Recognition System with Sparsely Labeled Instances Using Wi-Fi Signals

A Robust Sign Language Recognition System with Sparsely Labeled Instances Using Wi-Fi Signals A Robust Sign Language Recognition System with Sparsey Labeed Instances Using Wi-Fi Signas Jiacheng Shang, Jie Wu Center for Networked Computing Dept. of Computer and Info. Sciences Tempe University Motivation

More information

Relative Positioning from Model Indexing

Relative Positioning from Model Indexing Reative Positioning from Mode Indexing Stefan Carsson Computationa Vision and Active Perception Laboratory (CVAP)* Roya Institute of Technoogy (KTH), Stockhom, Sweden Abstract We show how to determine

More information

Resource Optimization to Provision a Virtual Private Network Using the Hose Model

Resource Optimization to Provision a Virtual Private Network Using the Hose Model Resource Optimization to Provision a Virtua Private Network Using the Hose Mode Monia Ghobadi, Sudhakar Ganti, Ghoamai C. Shoja University of Victoria, Victoria C, Canada V8W 3P6 e-mai: {monia, sganti,

More information

Proceedings of the International Conference on Systolic Arrays, San Diego, California, U.S.A., May 25-27, 1988 AN EFFICIENT ASYNCHRONOUS MULTIPLIER!

Proceedings of the International Conference on Systolic Arrays, San Diego, California, U.S.A., May 25-27, 1988 AN EFFICIENT ASYNCHRONOUS MULTIPLIER! [1,2] have, in theory, revoutionized cryptography. Unfortunatey, athough offer many advantages over conventiona and authentication), such cock synchronization in this appication due to the arge operand

More information

Grating cell operator features for oriented texture segmentation

Grating cell operator features for oriented texture segmentation Appeared in in: Proc. of the 14th Int. Conf. on Pattern Recognition, Brisbane, Austraia, August 16-20, 1998, pp.1010-1014. Grating ce operator features for oriented texture segmentation P. Kruizinga and

More information

Layer-Specific Adaptive Learning Rates for Deep Networks

Layer-Specific Adaptive Learning Rates for Deep Networks Layer-Specific Adaptive Learning Rates for Deep Networks arxiv:1510.04609v1 [cs.cv] 15 Oct 2015 Bharat Singh, Soham De, Yangmuzi Zhang, Thomas Godstein, and Gavin Tayor Department of Computer Science Department

More information

Absolute three-dimensional shape measurement with two-frequency square binary patterns

Absolute three-dimensional shape measurement with two-frequency square binary patterns 871 Vo. 56, No. 31 / November 1 217 / Appied Optics Research Artice Absoute three-dimensiona shape measurement with two-frequency square binary patterns CHUFAN JIANG AND SONG ZHANG* Schoo of Mechanica

More information

A Memory Grouping Method for Sharing Memory BIST Logic

A Memory Grouping Method for Sharing Memory BIST Logic A Memory Grouping Method for Sharing Memory BIST Logic Masahide Miyazai, Tomoazu Yoneda, and Hideo Fuiwara Graduate Schoo of Information Science, Nara Institute of Science and Technoogy (NAIST), 8916-5

More information

Neural Networks. Aarti Singh. Machine Learning Nov 3, Slides Courtesy: Tom Mitchell

Neural Networks. Aarti Singh. Machine Learning Nov 3, Slides Courtesy: Tom Mitchell Neura Networks Aarti Singh Machine Learning 10-601 Nov 3, 2011 Sides Courtesy: Tom Mitche 1 Logis0c Regression Assumes the foowing func1ona form for P(Y X): Logis1c func1on appied to a inear func1on of

More information

Crossing Minimization Problems of Drawing Bipartite Graphs in Two Clusters

Crossing Minimization Problems of Drawing Bipartite Graphs in Two Clusters Crossing Minimiation Probems o Drawing Bipartite Graphs in Two Custers Lanbo Zheng, Le Song, and Peter Eades Nationa ICT Austraia, and Schoo o Inormation Technoogies, University o Sydney,Austraia Emai:

More information

Digital Image Watermarking Algorithm Based on Fast Curvelet Transform

Digital Image Watermarking Algorithm Based on Fast Curvelet Transform J. Software Engineering & Appications, 010, 3, 939-943 doi:10.436/jsea.010.310111 Pubished Onine October 010 (http://www.scirp.org/journa/jsea) 939 igita Image Watermarking Agorithm Based on Fast Curveet

More information

Traffic Engineering for Provisioning Restorable Hose-Model VPNs

Traffic Engineering for Provisioning Restorable Hose-Model VPNs Traffic Engineering for Provisioning Restorabe Hose-Mode VPNs Yu-Liang Liu, Nonmember, Yeai S. Sun a) and Meng Chang Chen b), Member Summary Virtua Private Networks (VPNs) are overay networks estabished

More information

Approximate Closest Community Search in Networks

Approximate Closest Community Search in Networks Approximate Cosest Community Search in Networks Xin Huang, Laks V.S. Lakshmanan, Jeffrey Xu Yu, Hong Cheng University of British Coumbia, The Chinese University of Hong Kong {xin,aks}@cs.ubc.ca, {yu, hcheng}@se.cuhk.edu.hk

More information

Privacy Preserving Subgraph Matching on Large Graphs in Cloud

Privacy Preserving Subgraph Matching on Large Graphs in Cloud Privacy Preserving Subgraph Matching on Large Graphs in Coud Zhao Chang,#, Lei Zou, Feifei Li # Peing University, China; # University of Utah, USA; {changzhao,zouei}@pu.edu.cn; {zchang,ifeifei}@cs.utah.edu

More information

Stereo Matching with Energy Minimizing Snake Grid for 3D Face Modeling

Stereo Matching with Energy Minimizing Snake Grid for 3D Face Modeling Stereo Matching with Energy Minimizing Snake Grid for 3D Face Modeing Shafik Huq 1, Besma Abidi 1, Ardeshir Goshtasby 2, and Mongi Abidi 1 1 Imaging, Robotics, and Inteigent System (IRIS) Laboratory, Department

More information

Joint disparity and motion eld estimation in. stereoscopic image sequences. Ioannis Patras, Nikos Alvertos and Georgios Tziritas y.

Joint disparity and motion eld estimation in. stereoscopic image sequences. Ioannis Patras, Nikos Alvertos and Georgios Tziritas y. FORTH-ICS / TR-157 December 1995 Joint disparity and motion ed estimation in stereoscopic image sequences Ioannis Patras, Nikos Avertos and Georgios Tziritas y Abstract This work aims at determining four

More information

Utility-based Camera Assignment in a Video Network: A Game Theoretic Framework

Utility-based Camera Assignment in a Video Network: A Game Theoretic Framework This artice has been accepted for pubication in a future issue of this journa, but has not been fuy edited. Content may change prior to fina pubication. Y.LI AND B.BHANU CAMERA ASSIGNMENT: A GAME-THEORETIC

More information

A Top-to-Bottom View: Energy Analysis for Mobile Application Source Code

A Top-to-Bottom View: Energy Analysis for Mobile Application Source Code A Top-to-Bottom View: Energy Anaysis for Mobie Appication Source Code Xueiang Li John P. Gaagher Roskide University Emai: {xueiang, jpg}@ruc.dk arxiv:1510.04165v1 [cs.oh] 14 Oct 2015 Abstract Energy efficiency

More information

Service Chain (SC) Mapping with Multiple SC Instances in a Wide Area Network

Service Chain (SC) Mapping with Multiple SC Instances in a Wide Area Network Service Chain (SC) Mapping with Mutipe SC Instances in a Wide Area Network This is a preprint eectronic version of the artice submitted to IEEE GobeCom 2017 Abhishek Gupta, Brigitte Jaumard, Massimo Tornatore,

More information

A Comparison of a Second-Order versus a Fourth- Order Laplacian Operator in the Multigrid Algorithm

A Comparison of a Second-Order versus a Fourth- Order Laplacian Operator in the Multigrid Algorithm A Comparison of a Second-Order versus a Fourth- Order Lapacian Operator in the Mutigrid Agorithm Kaushik Datta (kdatta@cs.berkeey.edu Math Project May 9, 003 Abstract In this paper, the mutigrid agorithm

More information

Arithmetic Coding. Prof. Ja-Ling Wu. Department of Computer Science and Information Engineering National Taiwan University

Arithmetic Coding. Prof. Ja-Ling Wu. Department of Computer Science and Information Engineering National Taiwan University Arithmetic Coding Prof. Ja-Ling Wu Department of Computer Science and Information Engineering Nationa Taiwan University F(X) Shannon-Fano-Eias Coding W..o.g. we can take X={,,,m}. Assume p()>0 for a. The

More information

LEARNING causal structures is one of the main problems

LEARNING causal structures is one of the main problems cupc: CUDA-based Parae PC Agorithm for Causa Structure Learning on PU Behrooz Zare, Foad Jafarinejad, Matin Hashemi, and Saber Saehkaeybar arxiv:8.89v [cs.dc] Dec 8 Abstract The main goa in many fieds

More information

Load Balancing by MPLS in Differentiated Services Networks

Load Balancing by MPLS in Differentiated Services Networks Load Baancing by MPLS in Differentiated Services Networks Riikka Susitaiva, Jorma Virtamo, and Samui Aato Networking Laboratory, Hesinki University of Technoogy P.O.Box 3000, FIN-02015 HUT, Finand {riikka.susitaiva,

More information

Evolving Role Definitions Through Permission Invocation Patterns

Evolving Role Definitions Through Permission Invocation Patterns Evoving Roe Definitions Through Permission Invocation Patterns Wen Zhang EECS Dept. Vanderbit University Nashvie, TN, USA wen.zhang.1@vanderbit.edu David Liebovitz Dept. of Medicine Northwestern University

More information

A study of comparative evaluation of methods for image processing using color features

A study of comparative evaluation of methods for image processing using color features A study of comparative evauation of methods for image processing using coor features FLORENTINA MAGDA ENESCU,CAZACU DUMITRU Department Eectronics, Computers and Eectrica Engineering University Pitești

More information