By Topic

Computational Intelligence and Security, 2008. CIS '08. International Conference on

Date 13-17 Dec. 2008

Go

Filter Results

Displaying Results 1 - 25 of 113
  • [Front cover - Workshop]

    Publication Year: 2008 , Page(s): C1
    Save to Project icon | Request Permissions | PDF file iconPDF (67 KB)  
    Freely Available from IEEE
  • [Title page i - Workshop]

    Publication Year: 2008 , Page(s): i
    Save to Project icon | Request Permissions | PDF file iconPDF (18 KB)  
    Freely Available from IEEE
  • [Title page iii - Workshop]

    Publication Year: 2008 , Page(s): iii
    Save to Project icon | Request Permissions | PDF file iconPDF (48 KB)  
    Freely Available from IEEE
  • [Copyright notice - Workshop]

    Publication Year: 2008 , Page(s): iv
    Save to Project icon | Request Permissions | PDF file iconPDF (45 KB)  
    Freely Available from IEEE
  • Table of contents - Workshop

    Publication Year: 2008 , Page(s): v - xii
    Save to Project icon | Request Permissions | PDF file iconPDF (116 KB)  
    Freely Available from IEEE
  • Preface to CIS 2008 Workshop

    Publication Year: 2008 , Page(s): xiii
    Save to Project icon | Request Permissions | PDF file iconPDF (88 KB) |  | HTML iconHTML  
    Freely Available from IEEE
  • CIS 2008 Committees - Workshop

    Publication Year: 2008 , Page(s): xiv
    Save to Project icon | Request Permissions | PDF file iconPDF (126 KB)  
    Freely Available from IEEE
  • CIS 2008 Program Committee - Workshop

    Publication Year: 2008 , Page(s): xv - xvi
    Save to Project icon | Request Permissions | PDF file iconPDF (119 KB)  
    Freely Available from IEEE
  • Chinese Named Entity Recognition with CRFs: Two Levels

    Publication Year: 2008 , Page(s): 1 - 6
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (418 KB)  

    Named entity recognition (NER) is one of the key techniques in natural language processing tasks such as information extraction, text summarization and so on. Chinese NER is more complicated and difficult than other languages because of its characteristics. This paper investigates Chinese named entity recognition based on CRFs, and implements three main named entities, person, location, and organization recognition in two levels: word level and character level. Experiments are made to compare the two level models¿ performances. In the experiments, different training scales and feature sets are utilized to look into the models¿ relationships with training corpus and their ability in making use of different features. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Content Semantic Similarity Boosted Collaborative Filtering

    Publication Year: 2008 , Page(s): 7 - 11
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (227 KB) |  | HTML iconHTML  

    Collaborative filtering (CF) is one of the most promising techniques in recommender systems, providing personalized recommendations to users based on their previously expressed preferences in the form of ratings and those of other similar users. In practice, a large number of ratings from similar users are not available, due to the sparsity inherent to rating data. Consequently, recommendation quality can be poor. In this paper, we present an effective content semantic similarity boosted CF framework (CSSCF) for combining content meaning and collaboration. Our approach uses a content semantic similarity based rater (CSSR) to enhance existing user data, and then provides personalized suggestions through collaborative filtering. The new model is however more robust to data sparsity, because missing ratings are rated using the CSSR in advance. Experiments demonstrate that the proposed method gives better recommendations than pure collaborative filter. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Expansion Complex Fast-ICA Algorithm Based on Complex Orthogonal Space

    Publication Year: 2008 , Page(s): 12 - 15
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (108 KB) |  | HTML iconHTML  

    Based on independent component analysis (ICA), this paper discusses the complex signal model of blind source separation. With discussions of characteristics of complex-valued signals, paper establishes mixed and separated model of complex-valued signals used to define an orthogonal decomposition of the complex-valued space. Using the relative gradients method, based on 4-th order cumulant, it sets up a successive independent component recovery fixed-point algorithm of blind separation of complex-valued sources. Finally, under the framework of this discussion, it generates randomly several sub-Gaussian and super-Gaussian signals in order to verify the claims of effectiveness and feasibility, using the algorithm to carry out computer simulation experiments, and analyzes the results. The results shows that the method is very effective. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Novel Blind Image Watermarking Scheme Based on Support Vector Machine in DCT Domain

    Publication Year: 2008 , Page(s): 16 - 20
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (593 KB) |  | HTML iconHTML  

    This paper proposed a novel blind image watermark scheme in discrete cosine transform (DCT) domain. The original image is divided into small image blocks with size 8*8 first, and then the binary watermark bits are embedded into DCT domain of image blocks adaptively through a novel watermark embedding algorithm. The original image is not needed in the extraction algorithm. For gaining good fidelity and robustness, support vector machine (SVM) is adopted to simulate human visual system (HVS) and class image blocks into several types. Different embedding strengths are determined for different image types as human has different sensitivity to them. There characteristics of image blocks are used for classification. The embedding locations are selected randomly ranging from 10 to 24 of AC coefficients by a key to enhance the security of watermark. The experimental results show that our scheme has both good fidelity and strong robustness. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A New SVM Multi-Class Classification Method Based on Error-Correcting Code

    Publication Year: 2008 , Page(s): 21 - 24
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (209 KB) |  | HTML iconHTML  

    Traditional SVM (support vector machine) multi-class classification methods are mainly based on one-to-one and one-to-multi, which both have disadvantages in applications: slow computational speed and low classification precision. This paper introduces a new method based on error correcting code to reduce the training time and improve the classification precision. In view of the relations among the length, the Hsmming distance and the order of the code and the generalization ability of each SVM, we propose the principles of code table-designing and the center-range method that ascertains the code order to eliminate the problem caused by error correcting code in factual application. Finally the results of experiments of HRRP recognition show this improved method has high computational efficiency and batter generalization ability. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • cDNA Microarray Image Processing Using Spot Centroid of Intensity

    Publication Year: 2008 , Page(s): 25 - 29
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (978 KB) |  | HTML iconHTML  

    cDNA microarray is an independent platform that offers the ability to analyze large amount of data, and an important application in the organism¿s metabolism and the gene expression analysis. cDNA microarray Image analysis aims to measure the intensity of each spot in the scanned image and this intensity represents the amount of a specific gene of the studied cell. The result can directly affect the subsequent analysis such as identifying different genes with different expressions. microarray image analysis includes three tasks: spot gridding, segmentation and information extraction. In this paper, mathematical morphology and spot centroid of intensity are applied to achieve an automatic way of spot segmentation. Three analysis results, including this paper, ScanAlyze software, method of Hongwei Li and method of Angulo J are compared. It shows that the arithmetic of the paper is accurate, stable and effective. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Survey of Semi-Supervised Learning Methods

    Publication Year: 2008 , Page(s): 30 - 34
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (194 KB) |  | HTML iconHTML  

    In traditional machine learning approaches to classification, one uses only a labelled set to train the classifier. Labelled instances however are often difficult, expensive, or time consuming to obtain, as they require the efforts of experienced human annotators. Meanwhile unlabeled data may be relatively easy to collect, but there has been few ways to use them. Semi-supervised learning addresses this problem by using large amount of unlabeled data, together with the labelled data, to build better classifiers. Because semi-supervised learning requires less human effort and gives higher accuracy, it is of great interest both in theory and in practice. The paper discusses various important approaches to semi-supervised learning such as self-training, co-training(CO), expectation maximization (EM), CO-EM, Then how graph-based methods are useful is explained. All semi-supervised learning methods are classified into generative and discriminative methods. But experimental results show that the hybrid algorithm gives better classification accuracy. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Artificial Neural Network Based on CIEA

    Publication Year: 2008 , Page(s): 35 - 40
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (709 KB) |  | HTML iconHTML  

    Chaotic Immune evolutionary algorithm (CIEA) is proposed on the shortage of current algorithms training artificial neural network, biological immune mechanism and the characteristics of evolution. A novel artificial neural network based on chaotic immune evolutionary algorithm (CIEANN) is presented in this paper. The algorithm has the merits of chaos, immunity and evolutionary algorithm. It can ensure the ability of global search and local search and enhance the performances of the algorithm. At last, we analyzed the performance of CIEANN with a typical XOR problem and compared it with several common ANNs. The analysis results show that CIEANN converges quickly and avoids immature convergence and is an effective way to solve optimization problem. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Improved Particle Swarm Optimization Using Particle Reliving Strategy

    Publication Year: 2008 , Page(s): 41 - 43
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (187 KB) |  | HTML iconHTML  

    The particle swarm optimization (PSO) is one of the best efficient algorithms. It has obtained more and more attention and has been applied in many fields, such as machine design and circuit design. But it also has some disadvantages, such as prematurely and difficultly to convergence. To improvement the performance of PSO, particle reliving strategy is proposed. With this strategy, a criterion is used to judge whether the particle relives. If so, the particle will relive just like that when the algorithm initials. Some benchmark functions are used to illuminate that the successful probability of PSO is improved with particle reliving. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Particle Swarm Optimization Based on Number-Theoretical Net

    Publication Year: 2008 , Page(s): 44 - 47
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (192 KB) |  | HTML iconHTML  

    An improved particle swarm optimization algorithm was proposed to fit multi-peaks searching; this algorithm was combined with number-theoretical method. In this algorithm, Number-theoretic net was used to initialize the particles' position, and for the purpose of multi-peak searching, the evolution equation was modified. The result of PSO is fined by a method named creeping algorithm for improving convergence. The experimental results on several classical functions show that the improved algorithm can get the better results, and the results also indicate that the number theoretical net results the better ability of convergence because of its better randomicity. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Convergence Property Analysis for PSO Based on Cluster-Degree

    Publication Year: 2008 , Page(s): 48 - 51
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (318 KB) |  | HTML iconHTML  

    The concept of cluster-degree was put forward and distribute status of particle with different cluster-degree was studied. The reasonable parameters setting range based on cluster-degree was proposed. Under the direction of cluster-degree, many parameters which can get better searching results were found. So this paper is helpful for the choosing and adjustment of PSO parameters in practical application. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Applications of Particle Swarm Optimization and K-Nearest Neighbors to Emotion Recognition from Physiological Signals

    Publication Year: 2008 , Page(s): 52 - 56
    Cited by:  Papers (3)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (254 KB) |  | HTML iconHTML  

    Emotion recognition based on physiological signals has a significant future of research and applications. However, in the process of emotion recognition, it is difficult to obtain the most significant feature combinations. Dual-Structure Particle Swarm Optimization (DSPSO) is applied to select emotion features of physiological signals so as to improve the recognition rates in this paper. K-Nearest Neighbors (KNN) is applied to PSO to select optimal feature subsets. This paper proposed incremental K for avoiding indivisibility about multi-classification. In view of repeated emergence about same swarms when iteration tends to be convergent, look-up table method is presented to avoid superfluous calculation. The experiment results demonstrate that these improved methods are feasible and efficient. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An Improved MCPSO with Center Communication

    Publication Year: 2008 , Page(s): 57 - 61
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (249 KB) |  | HTML iconHTML  

    This paper proposes an improved multi-swarm cooperative particle swarm optimizer with center communication (MCPSO-CC) based on our previous proposed MCPSO algorithm, which enhances the particles based on the experience of master swarm and slave swarms. In our original MCPSO, there is no information sharing among slave swarms except that the information of the best performing particle is broadcasted to the master swarm. To deal with this issue in MCPSO-CC the population is divided into several identical sub-swarms and a center communication strategy is used to transfer the information among all the sub-swarms. To demonstrate the efficiency of the proposed MCPSO-CC algorithm, its performance is compared with SPSO and MCPSO on four well-know benchmark functions. Experimental results show that MCPSO-CC achieves not only better solutions but also faster convergence. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A Fast Algorithm to Find Optimal Double-Loop Networks

    Publication Year: 2008 , Page(s): 62 - 67
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (267 KB)  

    Double-loop networks have been widely studied as architecture for local area networks. A double-loop digraph G(N; s1, s2) has N vertices 0, 1,..., N-l and 2N edges of two types: s1-edge: i¿1 + s (mod N); i=0, 1,..., N-l, and s2-edge: i¿ s + s2 (modN); i=0, 1,..., N-l, for some fixed steps 1¿ s1<s2<N with gcd(N; s1, s2) = 1. Let D(N; s1, s2) be the diameter of G and let us define D(N) = min{ D(N; s1, s2) | 1¿s1<s2<N and gcd(N; s1, s2) = 1} and D(N) = min{D(N; 1, s)| 1<s<N}. Given a fixed number of vertices N, the general problem is to find steps s1 and s2, such that the digraph G(N; s1, s2) has minimum diameter D(N). A lower bound of this diameter is known to be lb(N)= [¿(3N)]-2. In this work, we give a simple and efficient algorithmic solution of the problem by using a geometrical approach. Given n, the algorithm has outputs s1, s2 and the minimum integer k=k(N), such that D(N; s1, s2 )=D(N)=lb(N)+k. The running time complexity of the algorithm is 0(k2 )0(N1/4 logN). Also, some flaws in the bibliography are corrected. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Learning through Decision Tree in Simulated Soccer Environment

    Publication Year: 2008 , Page(s): 68 - 70
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1134 KB) |  | HTML iconHTML  

    The robotic soccer is one of the most complex multiagent systems in which agents play the role of soccer players. The characteristics of such systems are: realtime, noisy, collaborative and adversarial. Therefore, playing agents must be capable to making decisions. This paper describes the use of decision tree to kick and catch the ball for two simulated soccer agents. One player shoots towards the goal and the other plays the role of goalkeeper. Experimental results have shown that rules achieved from decision tree lead to more effective operations in simulated soccer agent. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Research on the Business Operation Model Based on Semantic Web Service and MAS

    Publication Year: 2008 , Page(s): 71 - 75
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (208 KB) |  | HTML iconHTML  

    With Web service technology widely accepted, Web service has become the standard of resources package in the Internet. The approach of exchanging information and transaction is also from a single device to the collaboration in the worldwide network. To face the challenges posed by today¿s changing and uncertain business environment, traditional business system integration approaches are not sufficient anymore. This paper presents a model to system integration, which leverages Agent Technology and Semantic Web service Technology, especially their features to obtain agile business process behavior. This paper sketches the problem, analyzes the possibilities and advantages of the solution approach and describes the model in details. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Identification of Fuzzy Models Using Cartesian Genetic Programming

    Publication Year: 2008 , Page(s): 76 - 81
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (277 KB) |  | HTML iconHTML  

    Fuzzy models have capability for solving problem in different application such as pattern recognition, prediction and control. Nevertheless, it has to be emphasized that the identification of a fuzzy model is complex task with many local minima. Cartesian genetic programming provides a way to solve such complex optimization problem. In this paper, fuzzy model is in form of network. Cartesian genetic programming is used to optimize the antecedent part and recursive least square is used to optimized the consequent part. The initialization of membership function parameters are doing with fuzzy clustering. Benefit of the methodology is illustrated by simulation results. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.