Showing papers in "Cybernetics and Information Technologies in 2012"
TL;DR: An overview of the current methodology and practice in modeling and control of the grinding process in industrial ball mills is presented.
Abstract: The paper presents an overview of the current methodology and practice in modeling and control of the grinding process in industrial ball mills. Basic kinetic and energy models of the grinding process are described and the most commonly used control strategies are analyzed and discussed.
45 citations
TL;DR: The objective of this paper is to systematically analyze VSM, LSI and FCA for the task of IR using standard and real life datasets.
Abstract: Latent Semantic Indexing (LSI), a variant of classical Vector Space Model (VSM), is an Information Retrieval (IR) model that attempts to capture the latent semantic relationship between the data items. Mathematical lattices, under the framework of Formal Concept Analysis (FCA), represent conceptual hierarchies in data and retrieve the information. However, both LSI and FCA use the data represented in the form of matrices. The objective of this paper is to systematically analyze VSM, LSI and FCA for the task of IR using standard and real life datasets.
38 citations
TL;DR: An algorithm called Policy learning by Weighting Exploration with the Returns, or RL PoWER is reported to have learned the fastest gait through only physical experiments yet reported in the literature, 16.3% faster than reported for HyperNEAT.
Abstract: Legged robots are uniquely privileged over their wheeled counterparts in their potential to access rugged terrain. However, designing walking gaits by hand for legged robots is a difficult and time-consuming process, so we seek algorithms for learning such gaits to automatically using real world experimentation. Numerous previous studies have examined a variety of algorithms for learning gaits, using an assortment of different robots. It is often difficult to compare the algorithmic results from one study to the next, because the conditions and robots used vary. With this in mind, we have used an open-source, 3D printed quadruped robot called QuadraTot, so the results may be verified, and hopefully improved upon, by any group so desiring. Because many robots do not have accurate simulators, we test gait-learning algorithms entirely on the physical robot. Previous studies using the QuadraTot have compared parameterized splines, the HyperNEAT generative encoding and genetic algorithm. Among these, the research on the genetic algorithm was conducted by (G l e t t e et al., 2012) in a simulator and tested on a real robot. Here we compare these results to an algorithm called Policy learning by Weighting Exploration with the Returns, or RL PoWER. We report that this algorithm has learned the fastest gait through only physical experiments yet reported in the literature, 16.3% faster than reported for HyperNEAT. In addition, the learned gaits are less taxing on the robot and more repeatable than previous record-breaking gaits.
26 citations
TL;DR: This work considers the peculiar characteristics of reinforcement learning in robotics, and devise a combination of two algorithms from the literature of derivative-free optimization, which is well suited for robotics, as it involves both off-line learning in simulation and on- line learning in the real environment.
Abstract: We consider the problem of optimization in policy space for reinforcement learning. While a plethora of methods have been applied to this problem, only a narrow category of them proved feasible in robotics. We consider the peculiar characteristics of reinforcement learning in robotics, and devise a combination of two algorithms from the literature of derivative-free optimization. The proposed combination is well suited for robotics, as it involves both off-line learning in simulation and on-line learning in the real environment. We demonstrate our approach on a real-world task, where an Autonomous Underwater Vehicle has to survey a target area under potentially unknown environment conditions. We start from a given controller, which can perform the task under foreseeable conditions, and make it adaptive to the actual environment.
19 citations
TL;DR: A 15 degrees of Freedom dynamic model of a compliant humanoid robot is used, combined with reinforcement learning to perform global search in the parameter space to produce stable gaits.
Abstract: In ZMP trajectory generation using simple models, often a considerable amount of trials and errors are involved to obtain locally stable gaits by manually tuning the gait parameters. In this paper a 15 degrees of Freedom dynamic model of a compliant humanoid robot is used, combined with reinforcement learning to perform global search in the parameter space to produce stable gaits. It is shown that for a given speed, multiple sets of parameters, namely step sizes and lateral sways, are obtained by the learning algorithm which can lead to stable walking. The resulting set of gaits can be further studied in terms of parameter sensitivity and also to include additional optimization criteria to narrow down the chosen walking trajectories for the humanoid robot.
18 citations
TL;DR: A model for summarization from large documents using a novel approach has been proposed by considering one of the South Indian regional languages (Kannada) and deals with a single document summarization based on statistical approach.
Abstract: Abstract The method for filtering information from large volumes of text is called Information Extraction. It is a limited task than understanding the full text. In full text understanding, we express in an explicit fashion about all the information in a given text. But, in Information Extraction, we delimit in advance, as part of the specification of the task and the semantic range of the result. Only extractive summarization method is considered and developed for the study. In this article a model for summarization from large documents using a novel approach has been proposed by considering one of the South Indian regional languages (Kannada). It deals with a single document summarization based on statistical approach. The purpose of summary of an article is to facilitate the quick and accurate identification of the topic of the published document. The objective is to save prospective readers’ time and effort in finding the useful information in a given huge article. Various analyses of results were also discussed by comparing it with the English language.
16 citations
TL;DR: A trainable multilayer feed forward neural network has been designed for the classification purposes and the result reveals that the proposed approach can classify with a good performance rate of 98%.
Abstract: Accurate classification of images is essential for the analysis of mammograms in computer aided diagnosis of breast cancer. We propose a new approach to classify mammogram images based on fractal features. Given a mammogram image, we first eliminate all the artifacts and extract the salient features such as Fractal Dimension (FD) and Fractal Signature (FS). These features provide good descriptive values of the region. Second, a trainable multilayer feed forward neural network has been designed for the classification purposes and we compared the classification test results with K-Means. The result reveals that the proposed approach can classify with a good performance rate of 98%.
14 citations
TL;DR: A methodology that aims to detect and diagnose faults, using thermographs approaches for the digital image processing technique, is described.
Abstract: The paper presents an overview of the image-processing techniques. The set of basic theoretical instruments includes methods of mathematical analysis, linear algebra, probability theory and mathematical statistics, theory of digital processing of one-dimensional and multidimensional signals, wavelet-transforms and theory of information. This paper describes a methodology that aims to detect and diagnose faults, using thermographs approaches for the digital image processing technique.
13 citations
TL;DR: A novel framework to recognize individuals from gait, which is a PCA based classifier and it achieved an average individual recognition rate of 94% through cross-validation.
Abstract: Abstract We propose a novel framework to recognize individuals from gait, in order to improve HRI. We collected the motion data of the torso from 13 persons’ gait, using 2 IMU sensors. We developed Feature Value Method which is a PCA based classifier and we achieved an average individual recognition rate of 94% through cross-validation.
13 citations
TL;DR: A model for prediction of the outcome indicators of e-Learning, based on Balanced ScoreCard by Neural Networks (NN) is proposed and the highest accuracy of prognosis is obtained applying the method of Optimal Brain Damage (OBD) over the nonlinear neural network.
Abstract: A model for prediction of the outcome indicators of e-Learning, based on Balanced ScoreCard (BSC) by Neural Networks (NN) is proposed. In the development of NN models the problem of a small sample size of the data arises. In order to reduce the number of variables and increase the examples of the training sample, preprocessing of the data with the help of the methods Interpolation and Principal Component Analysis (PCA) is performed. A method for optimizing the structure of the neural network is applied over linear and nonlinear neural network architectures. The highest accuracy of prognosis is obtained applying the method of Optimal Brain Damage (OBD) over the nonlinear neural network. The efficiency and applicability of the method suggested is proved by numerical experiments on the basis of real data.
11 citations
TL;DR: This paper proposes an approach using perceptual grouping via a graph cut and its combinations with Convolutional Neural Network to achieve improved segmentation of SEM images.
Abstract: Abstract Detecting the neural processes like axons and dendrites needs high quality SEM images. This paper proposes an approach using perceptual grouping via a graph cut and its combinations with Convolutional Neural Network (CNN) to achieve improved segmentation of SEM images. Experimental results demonstrate improved computational efficiency with linear running time.
TL;DR: Performance analysis, cost analysis and cost-performance ratio analysis are done by comparing different cluster configurations, and high security is provided at this point for data using AES algorithm and also a password protection key for privileged user’s access.
Abstract: Cloud computing is a model where software applications and computing resources are accessed over Internet with minimal cost and effort by interacting with the service provider. Along with these benefits there are also some significant security concerns that need to be addressed for handling sensitive data and critical applications. The simultaneous use of multiple clouds can provide several potential benefits, such as high availability, fault tolerance and reduced infrastructural cost. The model proposed which is the implementation of a secured multi-cloud virtual infrastructure consists of a grid engine on top of the multi-cloud infrastructure to distribute the task among the worker nodes that are supplied with various resources from different clouds to enhance cost efficiency of the infrastructure set up and also to implement high availability feature. The Oracle grid engine is used to schedule the jobs to the worker nodes (in-house and cloud). Worker nodes will be acting like listeners to receive the job from the oracle grid engine master node. High security is provided at this point for data using AES algorithm and also a password protection key for privileged user’s access. Performance analysis, cost analysis and cost-performance ratio analysis are done by comparing different cluster configurations.
TL;DR: This new approach to explore the applicability of fuzzy logic in multiagent systems for choosing the best bidding strategy in electronic auction can solve the task for multi criteria selection of bidding strategy.
Abstract: The goal of this work is to explore the applicability of fuzzy logic in multi- agent systems for choosing the best bidding strategy in electronic auction. To find the multi-criterion ordering, agents use a fuzzy algorithm ARAKRI2 with direct aggregation operators MaxMin and MinAvg. The key difference between this new approach and known from the literature solution FTNA is in the lack of weighted coefficients. Despite the difference both algorithms give results that are similar. Therefore, the proposed approach can successfully solve the task for multi criteria selection of bidding strategy.
TL;DR: A framework that a robot can use to complete the ordering of a set of objects and was tested on object completion tasks in which the objects varied by weight, compliance, and height.
Abstract: This paper describes a framework that a robot can use to complete the ordering of a set of objects. Given two sets of objects, an ordered set and an unordered set, the robot’s task is to select one object from the unordered set that best completes the ordering in the ordered set. In our experiments, the robot interacted with each object using a set of exploratory behaviors, while recording feedback from two sensory modalities (audio and proprioception). For each behavior and modality combination, the robot used the feedback sequence to estimate the perceptual distance for every pair of objects. The estimated object distance features were subsequently used to solve ordering tasks. The framework was tested on object completion tasks in which the objects varied by weight, compliance, and height. The robot was able to solve all of these tasks with a high degree of accuracy.
TL;DR: This paper proposes a novel multiple kernel learning approach to generate a synthetic training set which is larger than the original training set, and shows that SVM classifier trained with synthetic patterns has demonstrated superior performance over the traditional SVMclassifier.
Abstract: Support Vector Machines (SVMs) have gained prominence because of their high generalization ability for a wide range of applications. However, the size of the training data that it requires to achieve a commendable performance becomes extremely large with increasing dimensionality using RBF and polynomial kernels. Synthesizing new training patterns curbs this effect. In this paper, we propose a novel multiple kernel learning approach to generate a synthetic training set which is larger than the original training set. This method is evaluated on seven of the benchmark datasets and experimental studies showed that SVM classifier trained with synthetic patterns has demonstrated superior performance over the traditional SVM classifier.
TL;DR: An algorithm of processing linguistic assessments is given that leads to aggregated opinions and conclusions in the form of linguistic evaluation.
Abstract: A simple use of intuitionistic fuzzy sets to determine and process the information describing the level of acceptance of people in a social group (e.g., pupils in a school class) is presented in the paper. An algorithm of processing linguistic assessments is given that leads to aggregated opinions and conclusions in the form of linguistic evaluation.
TL;DR: An approach for Information Extraction from Patient Records (PRs) in Bulgarian using N-grams, collocations and words’ distances allows us to cope with this problem and to extract automatically the attribute-value pairs with relatively high precision.
Abstract: Abstract This paper presents an approach for Information Extraction (IE) from Patient Records (PRs) in Bulgarian. The specific terminology and lack of resources in electronic format are some of the obstacles that make the task of current patient status data extraction in a structured format quite challenging. The usage of N-grams, collocations and words’ distances allows us to cope with this problem and to extract automatically the attribute-value pairs with relatively high precision.
TL;DR: An approach to active learning facilitated by the use of semantic technologies to shape the functionality of experimental Technology Enhanced Learning (TEL) environment with a built-in domain and pedagogical knowledge is presented.
Abstract: The paper presents an approach to active learning facilitated by the use of semantic technologies Some features of active learning and understanding of learning in humanities are discussed The specifics of a well defined learning task – learner’s authoring of analytical materials, grounded by materials from a digital library – are analyzed to shape the functionality of experimental Technology Enhanced Learning (TEL) environment with a built-in domain and pedagogical knowledge The environment structure and realization are discussed and a learning example is presented
TL;DR: The concept of Novel index tree (a variant of K-d tree) clubbed with K-Nearest Neighbor algorithm is proposed for efficient classification of data, as well as outliers and the concept of insurance dynamics is suggested for analyzing customer behavioral patterns.
Abstract: Abstract Extraction of customer behavioral patterns is a complex task and widely studied for various industrial applications under different heading viz., customer retention management, business intelligence and data mining. In this paper, authors experimented to extract the behavioral patterns for customer retention in Health care insurance. Initially, the customers are classified into three general categories - stable, unstable and oscillatory. To extract the patterns the concept of Novel index tree (a variant of K-d tree) clubbed with K-Nearest Neighbor algorithm is proposed for efficient classification of data, as well as outliers and the concept of insurance dynamics is proposed for analyzing customer behavioral patterns
TL;DR: A new model which combines the inverted pendulum approach with a three-dimensional (Cartesian) compliant model at the level of the center of mass is proposed, based on some assumptions that reduce the complexity but at the same time affect the precision.
Abstract: COMAN is a compliant humanoid robot. The introduction of passive compliance in some of its joints affects the dynamics of the whole system. Unlike traditional stiff robots, there is a deflection of the joint angle with respect to the desired one whenever an external torque is applied. Following a bottom up approach, the dynamic equations of the joints are defined first. Then, a new model which combines the inverted pendulum approach with a three-dimensional (Cartesian) compliant model at the level of the center of mass is proposed. This compact model is based on some assumptions that reduce the complexity but at the same time affect the precision. To address this problem, additional parameters are inserted in the model equation and an optimization procedure is performed using reinforcement learning. The optimized model is experimentally validated on the COMAN robot using several ZMP-based walking gaits.
TL;DR: A MultiObjective Genetic Modified Algorithm (MOGMA) is proposed, which is an improvement of the existing algorithm and a Pareto based fitness assignment is used in a multiobjective optimization problem.
Abstract: Multiobjective optimization based on genetic algorithms and Pareto based approaches in solving multiobjective optimization problems is discussed in the paper. A Pareto based fitness assignment is used − non-dominated ranking and movement of a population towards the Pareto front in a multiobjective optimization problem. A MultiObjective Genetic Modified Algorithm (MOGMA) is proposed, which is an improvement of the existing algorithm.
TL;DR: In this paper, new information inequalities involving f-divergences have been established using the convexity arguments and some well known inequalities such as the Jensen inequality and the Arithmetic-Geometric Mean (AGM) inequality.
Abstract: New information inequalities involving f-divergences have been established using the convexity arguments and some well known inequalities such as the Jensen inequality and the Arithmetic-Geometric Mean (AGM) inequality. Some particular cases have also been discussed.
TL;DR: A semantic technology based environment intended for developing technology enhanced learning applications in humanitarian problem domains and a preliminary evaluation of the environment based on an exploitation is presented.
Abstract: The paper describes a semantic technology based environment intended for developing technology enhanced learning applications in humanitarian problem domains. The environment consists of three layers: the storage layer contains heterogeneous repositories storing domain and pedagogical knowledge; the tool level contains a set of tools for processing different types of knowledge and the middleware layer is implemented as an extended search engine carrying out all necessary communications between the tools and the repositories. Some implementation issues are discussed and a preliminary evaluation of the environment based on an exploitation is presented.
TL;DR: The main advantage of the algorithm is that the derived general error covariance matrix equation is the same as this in the recursive least square algorithm and most of the well known RLS modifications for the tracking timevariant parameters can be directly implemented.
Abstract: An algorithm for multiple model adaptive control of a time-variant plant in the presence of measurement noise is proposed. This algorithm controls the plant using a bank of PID controllers designed on the base of time invariant input/output models. The control signal is formed as weighting sum of the control signals of local PID controllers. The main contribution of the paper is the objective function minimized to determine the weighting coefficients. The proposed algorithm minimizes the sum of the square general error between the model bank output and the plant output. An equation for on-line determination of the weighting coefficients is obtained. They are determined by the current value of the general error covariance matrix. The main advantage of the algorithm is that the derived general error covariance matrix equation is the same as this in the recursive least square algorithm. Thus, most of the well known RLS modifications for the tracking timevariant parameters can be directly implemented. The algorithm performance is tested by simulation. Results with both SISO and MIMO time variant plants are obtained.
TL;DR: New applications of the Ontology-to-Text Relation Strategy to Bulgarian Iconographic Domain shows that this strategy is good enough for more precise, but shallow results, and can be supported further by deep parsing techniques.
Abstract: Abstract The paper presents new applications of the Ontology-to-Text Relation Strategy to Bulgarian Iconographic Domain. First the strategy itself is discussed within the triple ontology-terminological lexicon-annotation grammars, then - the related works. Also, the specifics of the semantic annotation and evaluation over iconographic data are presented. A family of domain ontologies over the iconographic domain are created and used. The evaluation against a gold standard shows that this strategy is good enough for more precise, but shallow results, and can be supported further by deep parsing techniques.
TL;DR: Applying the considered approach, tight first order perturbation bounds are able to be computed for the LMIs’ solutions to the H∞ quadratic stability problem for discrete-time descriptor systems.
Abstract: Abstract In this paper we propose an approach to obtain first order perturbation bounds for the discrete-time Linear Matrix Inequalities (LMI) based H∞ quadratic stability problem for descriptor systems. Applying the considered approach we are able to compute tight first order perturbation bounds for the LMIs’ solutions to the H∞ quadratic stability problem for discrete-time descriptor systems. In the paper we present an approach to compute the estimates of the individual condition numbers of the considered LMIs. To illustrate the performance and applicability of the results obtained we present a numerical example
TL;DR: A novel robotics interface with perception and projection capabilities for facilitating the skill transfer process, which aims at allowing humans and robots to interact with each other in the same environment, with respect to visual feedback.
Abstract: In this work we discuss a novel robotics interface with perception and projection capabilities for facilitating the skill transfer process. The interface aims at allowing humans and robots to interact with each other in the same environment, with respect to visual feedback. During the learning process, the real workspace can be used as a graphical interface for helping the user to better understand what the robot has learned up to then, to display information about the task or to get feedback and guidance. Thus, the user can incrementally visualize and assess the learner's state and, at the same time, focus on the skill transfer without disrupting the continuity of the teaching interaction. We also propose a proof-of-concept, as a core element of the architecture, based on an experimental setting where a pico- projector and an rgb-depth sensor are mounted onto the end-effector of a 7-DOF robotic arm.
TL;DR: This paper proposes an algorithm that enables a learner to generalize a task representation from a small number of demonstrations of the same task and describes the supporting representation that is used in order to encode the generalized representation.
Abstract: Learning by demonstration is a natural approach that can be used to build a robot’s task repertoire. In this paper we propose an algorithm that enables a learner to generalize a task representation from a small number of demonstrations of the same task. The algorithm can generalize a wide range of situations that typically occur in daily tasks. The paper also describes the supporting representation that we use in order to encode the generalized representation. The approach is validated with experimental results on a broad range of generalizations.
TL;DR: The capabilities for open access to resource management in multimedia networks through Parlay X Web Services through Policy and Charging Control functions defined for evolved packet systems are investigated.
Abstract: Abstract The paper investigates the capabilities for open access to resource management in multimedia networks through Parlay X Web Services. Resource management allows the network operator to provide Quality of Service (QoS) to user sessions and to apply advanced charging. The study is based on the analysis of Policy and Charging Control (PCC) functions defined for evolved packet systems The PCC comprises flow-based charging including charging control and online credit control, gating control, and Quality of Service Control (QoS) control. The required functionality for open access to QoS management and advanced charging is identified. Parlay X Web Services are evaluated for the support of PCC and some enhancements are suggested. Implementation aspects are discussed and Parlay X interfaces are mapped on IMS control protocols. Use Cases of Parlay X Web services for PCC are presented.