Context Switching in a Run-Time Reconfigurable System

A distinguishing feature of reconfigurable computing over rapid prototyping is its ability to configure the computational fabric on-line while an application is running. Conventional reconfigurable computing platforms utilize commodity FPGAs, which typically have relatively long configuration times. Shrinking the configuration time down to the nanosecond region opens possibilities for rapid context switching and virtualizing the computational resources. An experimental context-switching FPGA, called the CSRC, has been created by BAE Systems, and gives researchers the opportunity to explore context-switching applications. This paper presents results obtained from constructing both control-driven and data-driven context switching applications on the CSRC device, along with unique properties of the run-time and compile-time environment.     

Hybrid accident simulation methodology using artificial neural networks for nuclear power plants

A hybrid accident simulation methodology for nuclear power plants is proposed to enhance the capabilities of compact simulator by introducing artificial neural networks. Two neural networks are trained with the target values obtained from the analyses of detailed computer codes and trained results are combined with the compact simulator to perform the following roles: (i) compensation for inaccuracies of a compact simulator occurring from simplified governing equation and reduced number of physical control volumes, and (ii) prediction of the critical parameter usually calculated from the sophisticated computer code: the autoassociative neural network improves the computational results of the compact simulator up to the accuracy level of detailed best estimate computer code, while the backpropagation neural network predicts the minimum departure from nucleate boiling ratio (DNBR). Simulations are carried out to verify the applicability of the proposed methodology for the loss of flow accidents and the results show that the neural networks can be used as a complementary tool to improve the results of a compact simulator.     

An agent for selecting optimal order set in EC marketplace

For most small manufacturing companies, the selection and completion of orders placed by buyers are closely linked to the load status of their production lines. The decision to accept an order, or the selection of optimal order set, critically depends on the production schedule when orders exceed production capacity. However, production scheduling is mainly performed by human experts, and small companies lack resources to implement such a task. As a result, most small manufacturers suffer from the difficulty to meet due dates or to make proper decision in accepting new orders. To address this problem, this paper develops an automatic agent that selects an optimal set of orders using commonly available Internet technology. The main engine of the selection agent is based on a typical job shop scheduling model, formulating and implementing it as an Integer Program (IP) model. We also translate IP into Genetic Algorithm (GA) to address its NP-hard problem. We conclude with a suggestion for an agent architecture that tackles Web-based order selection problems.     

Influence of calcination temperature on structural and optical properties of TiO2-SiO2 thin films prepared by sol-gel dip coating

We prepared TiO₂-SiO₂ thin films with various TiO₂/SiO₂ ratios by sol-gel dip coating method and explored the dependence of their structural and optical properties on calcination temperature. The absorption peaks relevant to Si—O, Si—O—Ti and Ti—O bonds appeared in the FTIR spectra. With increasing TiO₂ content, the intensity of Si—O bond peaks decreases and that of Ti—O bond peaks increases. The XRD results show that the temperature of transformation from amorphous to anatase phase is lowered as TiO₂ content increases. The crystallite size of anatase phase in composite thin films increases with increasing TiO₂ content and calcination temperature. At 1000°C, the mixed phase of anatase and rutile appears in the pure TiO₂ thin films. The rutile films are denser than the anatase films. The increase in refractive index of composite thin films with calcination temperature is related to the decreased thickness and increased density as a result of evaporation of water and organic matters below 400°C. On the other hand, it is related to the change in the crystal phase and crystallite size of the films over 400°C.     

Capturing Attention When Attention "Blinks".

Four experiments addressed the question of whether attention may be captured when the visual system is in the midst of an attentional blink (AB). Participants identified 2 target letters embedded among distractor letters in a rapid serial visual presentation sequence. In some trials, a square frame was inserted between the targets; as the only geometric object in the sequence, it constituted a singleton. Capture effects obtained when the AB was most severe and when it was over were compared. There were 3 main results. First, capture occurred even when the AB was crippling, suggesting that a singleton exogenously engaged attention even when processing of a previous target was continuing apace. Second, when the singleton contained the key target feature, capture effects were clearly manifest. Third, even when the singleton did not possess the key target feature, it still succeeded in capturing attention, although the effects were both feeble and fleeting. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A Personalized Defection Detection and Prevention Procedure based on the Self-Organizing Map and Association Rule Mining: Applied to Online Game Site

Customer retention is an increasinglypressing issue in today's competitiveenvironment. This paper proposes a personalizeddefection detection and prevention procedurebased on the observation that potentialdefectors have a tendency to take a couple ofmonths or weeks to gradually change theirbehaviour (i.e., trim-out their usage volume)before their eventual withdrawal. For thispurpose, we suggest a SOM (Self-Organizing Map)based procedure to determine the possiblestates of customer behaviour from pastbehaviour data. Based on this staterepresentation, potential defectors aredetected by comparing their monitoredtrajectories of behaviour states with frequentand confident trajectories of past defectors.Also, the proposed procedure is extended toprevent the defection of potential defectors byrecommending the desirable behaviour state forthe next period so as to lower the likelihoodof defection. For the evaluation of theproposed procedure, a case study has beenconducted for a Korean online game site. Theresult demonstrates that the proposed procedureis effective for defection prevention andefficiently detects potential defectors withoutdeterioration of prediction accuracy whencompared to that of the MLP (Multi-LayerPerceptron) neural networks.     

A nonlinear minimization approach to multiobjective and structured controls for discrete‐time systems

In this paper, a nonlinear minimization approach is proposed for multiobjective and structured controls for discrete‐time systems. The problem of finding multiobjective and structured controls for discrete‐time systems is represented as a quadratic matrix inequality problem. It is shown that the problem is reduced to a nonlinear minimization problem that has a concave objective function and linear matrix inequality constraints. An algorithm for the nonlinear minimization problem is proposed, which is easily implemented with existing semidefinite programming algorithms. The validity of the proposed algorithm is illustrated by comparisons with existing methods. In addition, applications of this work are demonstrated via numerical examples. Copyright © 2004 John Wiley & Sons, Ltd.     

Multitask bilateral learning for real‐time image enhancement

Nowadays, deep neural networks (DNNs) for image processing are becoming more complex; thus, reducing computational cost is increasingly important. This study highlights the construction of a DNN for real‐time image processing, training various image processing operators efficiently through multitask learning. For real‐time image processing, the proposed algorithm takes a joint upsampling approach through bilateral guided upsampling. For multitask learning, the overall network is based on an encoder‐decoder architecture, which consists of encoding, processing, and decoding components, in which the encoding and decoding components are shared by all the image processing operators. In the processing component, a semantic guidance map, which contains processing information for each image processing operator, is estimated using simple linear shifts of the shared deep features. Through these components, the proposed algorithm requires an increase of only 5% in the number of parameters to add another image processing operator and achieves faster and higher performance than that of deep‐learning‐based joint upsampling methods in local image processing as well as global image processing.