Many e-voting schemes have been proposed in the literature. However, none of them is both secure and practical. In this paper, a practical and secure electronic voting protocol for large-scale voting over the Internet is investigated. Blind signature is applied to a voter's ballot making it impossible for anyone to trace the ballot back to the voter. Unlike previous blind signature based schemes, in which the authority directly signs its blind signature on voters' ballots, the authority in the proposed scheme signs blind signature on the voter marks that are generated by voters from ballot serial numbers. Moreover, threshold cryptosystem has been used to guarantee the fairness of the voting process. Using blind signature, this scheme can support all types of election easily and flexibly. Since we haven’t use complex cryptographic techniques the proposed scheme is suitable for large scale elections.

Abstract. We study the problem of online coloring co-interval graphs. In this problem, a set of intervals on the real line is presented to the online algorithm in some arbitrary order, and the algorithm must assign each interval a color that is di®erent from the colors of all previously presented intervals not intersecting the current interval. It is known that the competitive ratio of the simple First-Fit algorithm on the class of co-interval graphs is at most 2.We show that for the class of unit co-interval graphs, where all intervals have equal length, the 2-bound on the competitive ratio of First-Fit is tight. On the other hand, we show that no deterministic online algorithm for coloring unit co-interval graphs can be better than 3/2-competitive. We then study the e®ect of randomization in our problem, and prove a lower bound of 4/3 on the competitive ratio of any randomized algorithm for the unit co-interval coloring problem. We also prove that for the class of general co-interval graphs no randomized algorithm has competitive ratio better than 3/2.

The main part of a machine vision system is to distinguish the object of interest (in the case of this project, a plant) from non-important regions (we refer it as background). Distinguishing the objects of interest is simplified if the high contrast between the objects of interest and background is created. The objective of this study is to find the best color index by which the algorithm is able to create the highest contrast between plant and non-plant regions. For this study, images were taken of varying numbers of wheat plants under several growth stages in a loamy sand soil and in diffused light condition. Three regions were predefined on the images; plant, pebble, and soil regions. Regions for plants, soil and pebbles were separately cropped within each image, aiming to provide a pooled representation for each object in each image. For each image, 13 mean color index were computed for each the three regions of interest (plant, soil, and pebble). The results of applying Analysis of variance (ANOVA) and consequently t-tests indicated that modified Excessive Green Index (MEGI) can potentially make the highest contrast between plant and non-plant regions rather than other color indices.

Continuous-State Reinforcement Learning (RL) has been recently favored because of the continuous nature of the real world RL problems and many theoretical approaches have been devised to handle the case. However, most of these methods presume that the structure of the agent's perceptual environment is fed to it. But this is not the case in many real situations. Inspired from the subjective view existing in the Cognitive Constructivist learning theory, in this paper, a new method is presented to discover and construct the structure of the environment in parallel with learning the optimal policy. To achieve these goals, the proposed approach incorporates the Bayesian formalism to organize the perceptual space while it tries to learn the optimal behavior using a Q-learning-like learning algorithm. These characteristics as a whole define a Reinforcement Learning algorithm which is developed based on a mixture of Cognitive Constructivism and traditional Behaviorism ideas. Simulation results demonstrate the viability and efficiency of the proposed algorithm on continuous state RL problems.

In several previous studies it has been shown that the generalization capabilities of humans through concept learning is reminiscent of Bayesian modeling. When discriminating concepts from one another, human subjects tend to focus on the relevant features of the subspace and ignore the irrelevant ones. In this paper we propose a Bayesian concept learning paradigm that utilizes unrestricted Bayesian networks to learn the required concepts for optimal decision making. This approach has several beneficial characteristics that a concept learning algorithm should hold. At first it can both learn form observing an expert performing the desired task and from its own experience while carrying it out. Secondly, it is a close and computationally feasible approximation to the Bayesian modeling capabilities of humans. Thirdly, the Markov blanket surrounding the decision variable can render the irrelevant features independent and therefore this approach can ignore them seamlessly from the feature subspace. The simulation and experimental results are promising and show that our approach can successfully extract the required temporally extended concepts for a mobile robot task.

This paper presents the theoretical development of nonlinear adaptive filter based on a concept of filtering in high dimensional space (HDS). The most common procedures for nonlinear estimation are the extended Kalman filter. The basic idea of the extended Kalman filter (EKF) is to linearize the state-space model at each time instant around the most recent state estimate. Once a linear model is obtained, the standard Kalman filter equations are applied. Main innovation in this paper is new linearization technique in EKF. The Linearization is performed by converting existing space to high dimensional space. HDS helps having linear space from nonlinear space. In this linear space, the standard Kalman filter gives rise to better results in estimation and prediction purposes. It is proven that MSE and error variance in this space is less than the input space. The proposed EKF is implemented in pedestrian tracking and results show that our method is superior to the standard extended Kalman filter.

In this paper, we propose a novel approach whereby a reinforcement learning agent attempts to understand its environment via meaningful temporally extended concepts in an unsupervised way. Our approach is inspired by findings in neuroscience on the role of mirror neurons in action-based abstraction. Since there are so many cases in which the best decision cannot be made just by using instant sensory data, in this study we seek to achieve a framework for learning temporally extended concepts from sequences of sensory-action data. To direct the agent to gather fertile information for concept learning, a reinforcement learning mechanism utilizing experience of the agent is proposed. Experimental results demonstrate the capability of the proposed approach in retrieving meaningful concepts from the environment. The concepts and the way of defining them are thought such that they not only can be applied to ease decision making but also can be utilized in other applications as elaborated in the paper.

Massively parallel systems are often composed of hundreds or thousands of components (such as routers, channels and connectors) that collectively possess failure rates higher than what arise in the ordinary systems. For these systems, new measures have been introduced that can evaluate the capability of a system for gracefully degradation. In the design of such systems, one of the most fundamental considerations is the reliability of their interconnected networks, which can be usually characterized by connectivity of the network topological structure. Resilience of graphs and various types of deterministic networks have attracted significant attention in the research literature. A classical problem in this line of study is to understand failure conditions which the network disconnects and/or starts to offer noticeably lower performance (such as increased routing distance) to its users. In this paper, we investigate the problem of network disconnection by means of simulation in the context of large-scale interconnect networks and understand how static patterns of node failure affect the resilience of such networks.

The design of a pattern classifier includes an attempt to select, among a set of possible features, a minimum subset of weakly correlated features that better discriminate the pattern classes. This is usually a difficult task in practice, normally requiring the application of heuristic knowledge about the specific problem domain. The selection and quality of the features representing each pattern have a considerable bearing on the success of subsequent pattern classification. Feature extraction is the process of deriving new features from the original features in order to reduce the cost of feature measurement, increase classifier efficiency, and allow higher classification accuracy. Many current feature extraction techniques involve linear transformations of the original pattern vectors to new vectors of lower dimensionality. While this is useful for data visualization and increasing classification efficiency, it does not necessarily reduce the number of features that must be measured since each new feature may be a linear combination of all of the features in the original pattern vector. In this paper a new approach is presented to feature extraction in which feature selection, feature extraction, and classifier training are performed simultaneously using a genetic algorithm. The genetic algorithm optimizes a vector of feature weights, which are used to scale the individual features in the original pattern vectors in either a linear or a nonlinear fashion.

Learning capabilities of an agent relies on the way that agent perceives the environment. When the agent’s sensations convey only partial information about the environment, there may be different situations that appear identical to the agent but require different actions to behave optimally. In this paper, we propose a new approach to improve XCS’s performance in Partially Observable Markov Decision Process (POMDP) using a newly introduced method to detect aliased states in the current environment. In our approach, at the initial state, there exists only a single main XCS which handles all of the environmental states. When an existing aliased state is detected using a simple mechanism, the system creates a new XCS, in addition to the main XCS which we call Cooperative XCS. The new XCS is responsible for handling this detected state. This mechanism allows the main XCS to handle non-aliased states and the other XCS’s cooperate with it by handling existing aliased states independently. Thus, the system is called Cooperative Specialized XCS and its performance is compared with some other classifier systems in some benchmark problems. The presented results demonstrate the effectiveness of our proposed approach.

Explore/Exploit dilemma is one of the most challenging issues in reinforcement learning area as well as learning classifier systems such as XCS. In this paper, an intelligent method is proposed to control the exploration rate in XCS to improve its long-term performance. This method is called Intelligent Exploration Method (IEM) and is applied to some benchmark problems to show the advantages of adaptive exploration rate for XCS.

Trust is the highest level layer of the Semantic Web. In this paper, a specialized social network concept - a Research Network - is defined formally, and a trust model is specified on top of it. The trust model specifies trust queries as well as an algorithm to answer them. Trust propagation, and the reductive method are discussed in detail. Also, the result of applying the trust query algorithm on a test data set is discussed. The trust model, as well as the related concepts have been defined formally.

Ensemble based methods have successfully been used in a variety of application. Recently using these methods in clustering algorithms has attracted a great deal of interest. Voting and averaging are two effective combining methods that are frequently used in multiple classifier systems. To use these methods in an unsupervised scenario (to combine multiple partitioning of data) the partitions must be relabeled first, i.e. similar partitions in different partitioning gets the same label. This phase has a great influence on the ensemble performance. In this paper a new heuristic label assignment method is proposed. The result of the Monte Carlo simulation and experimental results on real data show that the performance of ensemble method could be significantly improved using this method.

The visibility graph is a fundamental geometric structure which is useful in many applica- tions, including illumination and rendering, motion planning, pattern recognition, and sensor networks. While the concept of visibility graph is widely studied for 2D scenes, there is not any acceptable equivalence of visibility graph for 3D space. In this paper we explain some reason for this absence. Then we try to ¯nd a new way to de¯ne geometric structure in 3D space. Following our new way, we easily de¯ne a new structure called 3D visibility graph which we believe is the natural way to extend visibility graph in 3D scenes. We show how to compute it in an acceptable time. keywords: computational geometry, visibility graph, 3D visibility.

Exceptionally leaky transistors are increasingly more frequent in nano-scale technologies due to lower threshold voltage and its increased variation. Such leaky transistors may even change position with changes in the operating voltage and temperature, and hence, static physical redundancy is not sufficient to tolerate such threats to yield. We show that in SRAM cells this leakage depends on the cell value and propose a first softwarebased runtime technique that suppresses such abnormal leakages by storing safe values in the corresponding cache lines before going to standby mode. Analysis shows the performance penalty is, in the worst case, linearly dependent to the number of so-cured cache lines while the energy saving linearly increases by the time spent in standby mode. Analysis and experimental results on commercial processors confirm that the technique is viable if the standby duration is more than a small fraction of a second.

In recent years, the unit selection-based concatenative speech synthesis method using a large corpus has attracted great attention, as it produces more natural quality speech compared to the parameter driven models. Weights of cost functions of unit selection approach have great effect on output quality. Important proportion or weight of every feature must be determined such a manner that cost functions has suitable correlation by human perceptual. In this paper, we proposed a new approach to automatically determine optimal weights for target cost using classification and regression trees. In this method, an objective measure by suitable correlation to human perceptually is initially selected. So, for instances of every phoneme, a classification tree has build to predict objective measure. Therefore, the proportion importance of every feature in classifying data using regression trees are determined and considered as weight of this feature. The objective measure prediction has over 50% correlation using the proposed method that showed 65% improvement relation to previous methods.

This paper presents a report on research carried out in the field of networked collaborative learning. In particular, we present a theory/model-based approach applied to a distance education course that is developed and taught in a virtual learning environment. In this educational practice, our objective has been twofold: first, to improve distance teaching and learning, and second, to facilitate social interaction among students and between tutor and students via the Web. To that end, our research approach has been based on the following actions: First, we analysed the goals, needs, expectations and preferences of our students, based on a previous pilot experience on distance collaborative learning, in order to understand what is actually happening in networked learning when collaboration becomes an integrated part of the whole learning process. Second, we proceeded to the design, development and implementation of a new pedagogical practice, called Virtual Study Group, to encourage and enhance learning through collaborative construction of knowledge and reflective interaction which contribute to a deeper understanding of the course contents. The paper focuses mainly on the latter. Finally, we provide a critical analysis and evaluation of the outcomes of this experience and of the many issues arising from applying this collaborative pedagogical practice to a virtual learning environment.

We believe that future network applications will benefit by adopting key biological principles and mechanisms. This paper is a successful attempt in order to design a congestion control mechanism. This attempt is based on predator-prey mathematical model. We show that interaction of those Internet entities that involved in congestion control mechanisms is similar to predator-prey interaction. This similarity motivates us to map the predator-prey approach to the Internet congestion control mechanism and design a bio-inspired congestion control scheme. The results show that using appropriately defined parameters, this model leads to a stable, fair and high performance congestion control algorithm.

Mobile Ad-hoc Networking (MANET) is an emerging issue in research and industry since its inherent properties satisfy a broad range of requirements and dreams of the future networking. The infrastructure-less nature of MANETs as well as mobility requirement, makes the routing process as a major basis and consequently challenge for building such networks. Due to frequent changes in the topology of a MANET, route discovery process should be invoked frequently, resulting in significant time, processing, and energy overhead. One might mention that discovering and maintaining multiple routes might degrade the problem in terms of time complexity and energy consumption. In this paper, we discuss the design issues and challenges for a typical multipath routing algorithm for MANETs and identify the major ideas behind different approaches. We also provide a comparative survey of proposed methods as well as their appropriate design issues.

From high level synthesis point of view, target design can be divided into two parts: controller and datapath. Single shared bus is a suitable structure for datapath synthesis regarding interconnections costs which suffers from several drawbacks such as its low data communication bandwidth. It has also been shown that the wordlength of functional units has a great impact on design costs. A combination of both methods is the core idea of this paper which is offering an improved communication structure. In this method datapath is partitioned into groups connected to segmented shared buses and every partition has a different width and all the functional units connected to a bus partition have the same input/output word-lengths. Having controlled the group binding and word-length of the functional units, as well as the other synthesis parameters, a high-level synthesis tool is introduced to implement DSP algorithms in digital hardware. The tool uses a multi-objective optimization genetic algorithm to minimize the circuit area, delay, power consumption and digital noise by selecting optimal grouping and word-length for the shared bus structure. Results demonstrate that savings can be made in the overall system costs by applying this method.