Showing papers in "Computing and Informatics \/ Computers and Artificial Intelligence in 2012"

Evaluation Measures for Text Summarization

Abstract: We explain the ideas of automatic text summarization approaches and the taxonomy of summary evaluation methods. Moreover, we propose a new evaluation measure for assessing the quality of a summary. The core of the measure is covered by Latent Semantic Analysis (LSA) which can capture the main topics of a document. The summarization systems are ranked according to the similarity of the main topics of their summaries and their reference documents. Results show a high correlation between human rankings and the LSA-based evaluation measure. The measure is designed to compare a summary with its full text. It can compare a summary with a human written abstract as well; however, in this case using a standard ROUGE measure gives more precise results. Nevertheless, if abstracts are not available for a given corpus, using the LSA-based measure is an appropriate choice.

A Neural Network Approach to Flood Mapping Using Satellite Imagery

Abstract: This paper presents a new approach to flood mapping using satellite synthetic-aperture radar (SAR) images that is based on intelligent techniques. In particular, we apply artificial neural networks, self-organizing Kohonen's maps (SOMs), for SAR image segmentation and classification. Our approach was used to process data from different satellite SAR instruments (ERS-2/SAR, ENVISAT/ASAR, RADARSAT-1) for different flood events: the Tisza river, Ukraine and Hungary, 2001; the Huaihe river, China, 2007; the Mekong river, Thailand and Laos, 2008; and the Koshi river, India and Nepal, 2008.

A Genetic Algorithm Approach for the Capacitated Single Allocation P-Hub Median Problem

Abstract: In this paper the Capacitated Single Allocation p-Hub Median Problem (CSApHMP) is considered. This problem has a wide range of applications within the design of telecommunication and transportation systems. A heuristic method, based on a genetic algorithm (GA) approach, is proposed for solving the CSApHMP. The described algorithm uses binary encoding and modified genetic operators. The caching technique is also implemented in the GA in order to improve its effectiveness. Computational experiments demonstrate that the GA method quickly reaches optimal solutions for hub instances with up to 50 nodes. The algorithm is also benchmarked on large scale hub instances with up to 200 nodes that are not solved to optimality so far.

Study and comparison of rule-based and statistical catalan-spanish machine translation systems

Abstract: Machine translation systems can be classified into rule-based and corpus-based approaches, in terms of their core methodology. Since both paradigms have been largely used during the last years, one of the aims in the research community is to know how these systems differ in terms of translation quality. To this end, this paper reports a study and comparison of several specific Catalan-Spanish machine translation systems: two rule-based and two corpus-based (particularly, statistical-based) systems, all of them freely available on the web. The translation quality analysis is performed under two different domains: journalistic and medical. The systems are evaluated by using standard automatic measures, as well as by native human evaluators. In addition to these traditional evaluation procedures, this paper reports a novel linguistic evaluation, which provides information about the errors encountered at the orthographic, morphological, lexical, semantic and syntactic levels. Results show that while rule-based systems provide a better performance at orthographic and morphological levels, statistical systems tend to commit less semantic errors. Furthermore, results show all the evaluations performed are characterised by some degree of correlation, and human evaluators tend to be specially critical with semantic and syntactic errors.

Some Applications of Spiking Neural P Systems

Abstract: In this paper we investigate some applications of spiking neural P systems regarding their capability to solve some classical computer science problems. In this respect versatility of such systems is studied to simulate a well known parallel computational model, namely the Boolean circuits. In addition, another notorious application -- sorting -- is considered within this framework.

An Efficient Genetic Algorithm for Solving the Multi-Level Uncapacitated Facility Location Problem

Abstract: In this paper a new evolutionary approach for solving the multi-level uncapacitated facility location problem (MLUFLP) is presented. Binary encoding scheme is used with appropriate objective function containing dynamic programming approach for finding sequence of located facilities on each level to satisfy clients' demands. The experiments were carried out on the modified standard single level facility location problem instances. Genetic algorithm (GA) reaches all known optimal solutions for smaller dimension instances, obtained by total enumeration and CPLEX solver. Moreover, all optimal/best known solutions were reached by genetic algorithm for a single-level variant of the problem.

Accurate Pupil Features Extraction Based on New Projection Function

Abstract: Accurate pupil features extraction is a key step for iris recognition. In this paper, we propose a new algorithm to extract pupil features precisely within gray level iris images. The angular integral projection function (AIPF) is developed as a general function to perform integral projection along angular directions, both the well known vertical and horizontal integral projection functions can be viewed as special cases of AIPF. Another implementation for AIPF based on localized Radon transform is also presented. First, the approximate position of pupil center is detected. Then, a set of pupil's radial boundary points are detected using AIPF. Finally, a circle to the detected boundary points is fitted. Experimental results on 2655 iris images from CASIA V3.0 show high accuracy with rapid execution time.

Analyzing Approaches to Collective Knowledge Recall

Abstract: People and organizations frequently need to recall past events that, for some reason, were not documented when they occurred. The successful reconstitution of past events depends on several variables, such as how long ago the event occurred, and whether key people are still available to tell what they know. Although it is sometimes difficult to restore all known events, an adequate recall process can get closer. This paper examines three knowledge recall methods and compares them in a set of controlled experiments. The group storytelling approach is used in two of the methods, one of which is supported by a groupware. The paper also evaluates the benefits and the drawbacks of using the group storytelling technology.

Hardware Accelerator of Cartesian Genetic Programming with Multiple Fitness Units

Abstract: A new accelerator of Cartesian genetic programming is presented in this paper. The accelerator is completely implemented in a single FPGA. The proposed architecture contains multiple instances of virtual reconfigurable circuit to evaluate several candidate solutions in parallel. An advanced memory organization was developed to achieve the maximum throughput of processing. The search algorithm is implemented using the on-chip PowerPC processor. In the benchmark problem (image filter evolution) the proposed platform provides a significant speedup (170) in comparison with a highly optimized software implementation. Moreover, the accelerator is 8 times faster than previous FPGA accelerators of image filter evolution.

Co-Evolutionary Multi-Agent System with Speciation and Resource Sharing Mechanisms

Abstract: Niching techniques for evolutionary algorithms are used in order to locate basins of attraction of the local minima of multi-modal fitness functions. Co-evolutionary techniques are aimed at overcoming limited adaptive capabilities of evolutionary algorithms resulting from the loss of useful population the idea of niching co-evolutionary multi-agent system (NCoEMAS)is introduced. In such a system the species formation phenomena occurs within one of the pre-existing species as a result of co-evolutionary interactions. The results of experiments with Rastrigin and Schwefel multi-modal test functions aimed at the comparison of NCoEMAS to other niching techniques are presented. Also, the resource sharing mechanism's parameters on the quality of speciation processes inNCoEMAS are investigated.

A Model of User Preference Learning for Content-Based Recommender Systems

Abstract: This paper focuses to a formal model of user preference learning for content-based recommender systems. First, some fundamental and special requirements to user preference learning are identified and proposed. Three learning tasks are introduced as the exact, the order preserving and the iterative user preference learning tasks. The first two tasks concern the situation where we have the user's rating available for a large part of objects. The third task does not require any prior knowledge about the user's ratings (i.e. the user's rating history). Local and global preferences are distinguished in the presented model. Methods for learning these preferences are discussed. Finally, experiments and future work will be described.

History-based Self-Organizing Traffic Lights

Abstract: Managing traffic in cities is nowadays a complex problem involving considerable physical and economical resources. Multi-agent Systems (MAS) consist of a set of distributed, usually co-operating, agents that act autonomously. The traffic in a city can be simulated by a MAS with different agents, cars and traffic lights, that interact to obtain an overall goal: to reduce average waiting times for the traffic users. In this paper, we describe an agent-based simulator to model traffic in cities. Using this simulator, we present a self-organizing solution to efficiently manage urban traffic. We compare our proposal with recent approaches, providing better results than classical and alternative self-organizing methods, with lower resources and investments.

A Comparative Study of Queue, Delay, and Loss Characteristics of AQM Schemes in QoS-enabled Networks

Abstract: One of the major component in a QoS-enabled network is active queue management (AQM). Over the last decade numerous AQM schemes have been proposed in the literature. However, much recent work has focused on improving AQM performance through alternate approaches. This study focuses on an unbiased comparative evaluation of the various proposals. The evaluation methodology adopted is the following: we first define the relationship between the terminologies used in this paper, briefly introduce the queue, delay, and loss characteristics-- a subset of network characteristics that can be used to describe the behavior of network entities, and give their mathematical description. Next, we present a method that would be a successful case study based on the NS simulation technique and simulation-based comparisons of AQM schemes chosen, which will help understand how they differ from in terms of per-node queueing information and per-flow end-to-end behavior. Simulation results showed that PI schemes, a feedback-based mechanism, can assist delay sensitive applications to adapt dynamically to underlying network and to stabilize the end-to-end QoS within an acceptable requirement. To understand this attribute and behavior is important for the proper design of queue disciplines, for the provisioning of queues and link capacity, and for choosing parameters in simulation.

Peer-to-Peer Networks and Computation: Current Trends and Future Perspectives

Abstract: This research papers examines the state-of-the-art in the area of P2P networks/computation. It attempts to identify the challenges that confront the community of P2P researchers and developers, which need to be addressed before the potential of P2P-based systems, can be effectively realized beyond content distribution and file-sharing applications to build real-world, intelligent and commercial software systems. Future perspectives and some thoughts on the evolution of P2P-based systems are also provided.

Automatic Control Synthesis for Agents and Their Cooperation in MAS

Abstract: Automatic synthesis of control for a kind of DES (discrete-event systems) is discussed and an approach to it is proposed and presented. The approach consists in the proposal of the control synthesis procedure based on bipartite directed graphs yielding both the feasible control trajectories and the corresponding state ones. Soundness of the approach is tested on examples. Then, the usage of the approach is combined with the supervisor synthesis in order to complement it. Applicability of such approach is demonstrated by means of several illustrative examples of both the single agents and the agent cooperation in MAS.

A Comparison of Decision Tree Classifiers for Automatic Diagnosis of Speech Recognition Errors.

Abstract: Present speech recognition systems are becoming more complex due to technology advances, optimizations and special requirements such as small computation and memory footprints. Proper handling of system failures can be seen as a kind of fault diagnosis. Motivated by the success of decision tree diagnosis in other scientific fields and by their successful application in speech recognition in the last decade, we contribute to the topic mainly in terms of comparison of different types of decision trees. Five styles are examined: CART (testing three different splitting criteria), C4.5, and then Minimum Message Length (MML), strict MML and Bayesian styles decision trees. We apply these techniques to data of computer speech recognition fed by intrinsically variable speech. We conclude that for this task, CART technique outperforms C4.5 in terms of better classification for ASR failures.

A Hybrid Test Optimization Framework - Coupling Genetic Algorithm with Local Search Technique

Abstract: Quality of test cases is determined by their ability to uncover as many errors as possible in the software code. In our approach, we applied Hybrid Genetic Algorithm (HGA) for improving the quality of test cases. This improvement can be achieved by analyzing both mutation score and path coverage of each test case. Our approach selects effective test cases that have higher mutation score and path coverage from a near infinite number of test cases. Hence, the final test set size is reduced which in turn reduces the total time needed in testing activity. In our proposed framework, we included two improvement heuristics, namely RemoveTop and LocalBest, to achieve near global optimal solution. Finally, we compared the efficiency of the test cases generated by our approach against the existing test case optimization approaches such as Simple Genetic Algorithm (SGA) and Bacteriologic Algorithm (BA) and concluded that our approach generates better quality test cases.

Approximability of the Minimum Steiner Cycle Problem

Abstract: In this paper, we consider variants of a new problem that we call minimum Steiner cycle problem (SCP). The problem is defined as follows. Given is a weighted complete graph and a set of terminal vertices. In the SCP problem, we are looking for a minimum-cost cycle that passes through every terminal exactly once and through every other vertex of the graph at most once. We show that, if P NP, there is no approximation algorithm for SCP on directed graphs with an approximation ratio polynomial in the input size. Moreover, this result holds even in the case when the number of terminals is 4. On the contrary, we show that SCP on undirected graphs with constant number of terminals and edge costs satisfying the beta-relaxed triangle inequality is approximable with the ratio beta^2+beta. When the number of terminals k is a part of the input, the problem is obviously a generalization of TSP. For the metric case, we present a 3/2- and a 2/3 log_2 k-approximation algorithm for undirected and directed graphs G=(V,E), respectively. For the case with the beta-relaxed triangle inequality, we present a (beta^2+beta)-approximation algorithm.

Evolutionary Approaches for Multi-Objective Next Release Problem

Abstract: In software industry, a common problem that the companies face is to decide what requirements should be implemented in the next release of the software. This paper aims to address the multi-objective next release problem using search based methods such as multi-objective evolutionary algorithms for empirical studies. In order to achieve the above goal, a requirement-dependency-based multi-objective next release model (MONRP/RD) is formulated firstly. The two objectives we are interested in are customers' satisfaction and requirement cost. A popular multi-objective evolutionary approach (MOEA), NSGA-II, is applied to provide the feasible solutions that balance between the two objectives aimed. The scalability of the formulated MONRP/RD and the influence of the requirement dependencies are investigated through simulations as well. This paper proposes an improved version of the multi-objective invasive weed optimization and compares it with various state-of-the-art multi-objective approaches on both synthetic and real-world data sets to find the most suitable algorithm for the problem.

Efficient mining of Fuzzy Association Rules from the Pre-Processed Dataset

Abstract: Association rule mining is an active data mining research area. Recent years have witnessed many efforts on discovering fuzzy associations. The key strength of fuzzy association rule mining is its completeness. This strength, however, comes with a major drawback to handle large datasets. It often produces a huge number of candidate itemsets. The huge number of candidate itemsets makes it ineffective for a data mining system to analyze them. In the end, it produces a huge number of fuzzy associations. This is particularly true for datasets whose attributes are highly correlated. The huge number of fuzzy associations makes it very difficult for a human user to analyze them. Existing research has shown that most of the discovered rules are actually redundant or insignificant. In this paper, we propose a novel technique to overcome these problems; we are preprocessing the data tuples by focusing on similar behaviour attributes and ontology. Finally, the efficiency and advantages of this algorithm have been proved by experimental results.

A Quantitative Characterization of Weighted Kripke Structures in Temporal Logic

Abstract: We extend the usual notion of Kripke structures with a weighted transition relation and generalize the classical Boolean interpretation of CTL to a map which assigns to states and temporal formulae a real-valued distance describing the degree of satisfaction. We describe a general approach to obtaining quantitative interpretations for a generic extension of the CTL syntax and show that, for one such interpretation, the logic is both adequate and expressive with respect to quantitative bisimulation.

Solving the Generalized Vertex Cover Problem by Genetic Algorithm

Abstract: In this paper an evolutionary approach to solving the generalized vertex cover problem (GVCP) is presented. Binary representation and standard genetic operators are used along with the appropriate objective function. The experiments were carried out on randomly generated instances with up to 500 vertices and 100000 edges. Performance of the genetic algorithm (GA) is compared with CPLEX solver and 2-approximation algorithm based on LP relaxation. The genetic algorithm outperformed both CPLEX solver and 2-approximation heuristic.

Agent-Based Fault Tolerant Distributed Event System

Abstract: In the last years, event-based communication style has been extensively studied and is considered a promising approach to develop large scale distributed systems. The historical development of event based systems has followed a line which has evolved from channel-based systems, to subject-based systems, next content-based systems and finally type-based systems which use objects as event messages. According to this historical development the next step should be usage of agents in event systems. In this paper, we propose a new model for Agent Based Distributed Event Systems, called ABDES, which combines the advantages of event-based communication and intelligent mobile agents into a flexible, extensible and fault tolerant distributed execution environment.

Nearest Neighbor Clustering over Partitioned Data

Abstract: Most clustering algorithms assume that all the relevant data are available on a single node of a computer network. In the emerging distributed and networked knowledge environments, databases relevant for computations may reside on a number of nodes connected by a communication network. These data resources cannot be moved to other network sites due to privacy, security, and size considerations. The desired global computation must be decomposed into local computations to match the distribution of data across the network. The capability to decompose computations must be general enough to handle different distributions of data and different participating nodes in each instance of the global computation. In this paper, we present a methodology and algorithm for clustering distributed data in d-dimensional space, using nearest neighbor clustering, wherein each distributed data source is represented by an agent. Each such agent has the capability to decompose global computations into local parts, for itself and for agents at other sites. The global computation is then performed by the agent either exchanging some minimal summaries with other agents or traveling to all the sites and performing local tasks that can be done at each local site. The objective is to perform global tasks with a minimum of communication or travel by participating agents across the network.

Management of data access with quality of service in pl-grid environment

Abstract: e-Science applications increasingly require both computational power and storage resources, currently supported with a certain level of quality. Since in the grid and cloud environments, where we can execute the e-Science applications, heterogeneity of storage systems is higher than that of computational power resources, optimization of data access defines one of challenging tasks nowadays. In this paper we present our approach to management of data access in the grid environment. The main issue is to organize data in such a way that users requirements in the form of QoS/SLA are met. For this purpose we make use of a storage monitoring system and a mass storage system model -- CMSSM. The experiments are performed in the PL-Grid environment.

Optimizing an LTS-Simulation Algorithm

Abstract: When comparing the fastest algorithm for computing the largest simulation preorder over Kripke structures with the one for labeled transition systems (LTS), there is a noticeable time and space complexity blow-up proportional to the size of the alphabet of an LTS. In this paper, we present optimizations that suppress this increase of complexity and may turn a large alphabet of an LTS to an advantage. Our experimental results show significant speed-ups and memory savings. Moreover, the optimized algorithm allows one to improve asymptotic complexity of procedures for computing simulations over tree automata using recently proposed algorithms based on computing simulation over certain special LTS derived from a tree automaton.

Security and QoS Integration Model for MANETs

Abstract: The new model used to integrating security and Quality of Service (QoS) as one parameter in mobile ad-hoc network (MANET) is introduced and studied in this article. Security and QoS represent a highly important field o research in MANET and they are still being considered separately with no mechanisms used to establish cooperation between them. This new model provides alternative to cooperation between QoS and security via cross layer design (CLD) and modified security service vector. Performance analysis of the new designed model is introduced too. It is also considered herein how processing of the new integrating model affects the performance of the MANET networks.

Optimization of Quadratic Assignment Problem Using Self Organising Migrating Algorithm

Abstract: This paper introduces the primary research involving Self Organinsing Migrating Algorithm (SOMA) to the permutative problem of Quadratic Assignment. SOMA is transformed from its canonical form to successfully solve permutative optimization problems. Conversion and repairment routines are added to the generic SOMA. The results presented outline the high effectiveness of SOMA for solving QAP problems.

Computing user reputation in a social network of web 2.0

Abstract: In the Web 2.0 era, people not only read web contents but create, upload, view, share and evaluate all contents on the web. This leads us to introduce a new type of social network based on user activity and content metadata. We notice that we can determine the quality of related contents using this new social network. Based on this observation, we introduce a user evaluation algorithm for user-generated video sharing website. First, we make a social network of users from video contents and related social activities such as subscription, uploading or favorite. We then use a modified PageRank algorithm to compute user reputation from the social network. We re-calculate the content scores using user reputations and compare the results with a standard BM25 result. We apply the proposed approach to YouTube and demonstrate that the user reputation is closely related to the number of subscriptions and the number of uploaded contents. Furthermore, we show that the new ranking results relied on the user reputation is better than the standard BM25 approach by experiments.

Semantic-oriented performance monitoring of distributed applications

Abstract: Monitoring services are an essential component of large-scale computing infrastructures due to providing information which can be used by humans as well as applications to closely follow the progress of computations, to evaluate the performance of ongoing computing, etc. However, the users are usually left alone with performance measurements as to the interpreting and detecting of execution flaws. In this paper we present an approach to the performance monitoring of distributed applications based on semantic information about the monitored objects involved in the application execution. This allows to automate the guidance on what to measure further to come to a source of performance flaws as well to enable reacting on interesting events, e.g. on exceeding SLA parameters. Our research comprises the implementation of a robust system with semantics, which is not biased to an underlying ``physical'' monitoring system, giving the end user the power of intelligent monitoring functionality as well as the independence of the heterogeneity of distributed infrastructures.