The science of making ERORS: what error tolerance implies forcapacity in neural networks

Discusses the development of formal protocols for handling error tolerance which allow a precise determination of the computational gains that may be expected. The error protocols are illustrated in the framework of a densely interconnected neural network architecture (with associative memory the putative application), and rigorous calculations of capacity ar shown. Explicit capacities are also derived for the case of feedforward neural network configurations     

Subject independent human action recognition using spatio-depth information and meta-cognitive RBF network

In this paper, we present a machine learning approach for subject independent human action recognition using depth camera, emphasizing the importance of depth in recognition of actions. The proposed approach uses the flow information of all 3 dimensions to classify an action. In our approach, we have obtained the 2-D optical flow and used it along with the depth image to obtain the depth flow (Z motion vectors). The obtained flow captures the dynamics of the actions in space–time. Feature vectors are obtained by averaging the 3-D motion over a grid laid over the silhouette in a hierarchical fashion. These hierarchical fine to coarse windows capture the motion dynamics of the object at various scales. The extracted features are used to train a Meta-cognitive Radial Basis Function Network (McRBFN) that uses a Projection Based Learning (PBL) algorithm, referred to as PBL-McRBFN, henceforth. PBL-McRBFN begins with zero hidden neurons and builds the network based on the best human learning strategy, namely, self-regulated learning in a meta-cognitive environment. When a sample is used for learning, PBL-McRBFN uses the sample overlapping conditions, and a projection based learning algorithm to estimate the parameters of the network. The performance of PBL-McRBFN is compared to that of a Support Vector Machine (SVM) and Extreme Learning Machine (ELM) classifiers with representation of every person and action in the training and testing datasets. Performance study shows that PBL-McRBFN outperforms these classifiers in recognizing actions in 3-D. Further, a subject-independent study is conducted by leave-one-subject-out strategy and its generalization performance is tested. It is observed from the subject-independent study that McRBFN is capable of generalizing actions accurately. The performance of the proposed approach is benchmarked with Video Analytics Lab (VAL) dataset and Berkeley Multi-modal Human Action Database (MHAD).     

Necessary and sufficient conditions for robust identification of uncertain LTI systems

Robust identification of uncertain systems arises whenever a chosen family of models does not completely describe reality. In these situations the issue of unmodeled dynamics gains significance in addition to random measurement noise. To deal with such mixed stochastic/deterministic settings we introduce a novel notion for robust consistency, which requires that the expectation (with respect to noise) of the worst-case (with respect to unmodeled dynamics) identification error asymptotically approach zero. It turns out that this notion leads to transparent necessary and sufficient conditions. We show that robust consistency holds, if and only if there is an instrument-input-pair capable of annihilating the residual error as well as stochastic noise. An extension of this result to the well-known “bounded but unknown” noise model shows that if we were to remove a set of Lebesgue measure zero, the error bound asymptotically approaches zero.     

混合信号FPGA的"智能"系统验证流程

为了满足市场对更高性能、更小体积及更低成本和功耗不断增长的需求,系统设计人员需要将更高级别的混合信号功能集成到系统级芯片(SoC)中。随着这些SoC设计中混合信号器件的增加,基本的功能验证在半导体的早期制备中变得十分重要。没有这种验证,系统设计人员将需要为半导体制备     

Architecture, Infrastructure, and Broadband Civic Network Design: An Institutional View

Cultural values frame architectures, and architectures motivate infrastructures-by which we mean the foundational telecommunications and Internet access services that software applications depend on. Design is the social process that realizes architectural elements in an infrastructure. This process is often a conflicted one where transformative visions confront the realities of entrenched power, where innovation confronts pressure from institutionalized interests and practices working to resist change and reproduce the status quo in the design outcome. We use this viewpoint to discuss design aspects of the Urban-net, a broadband civic networking case. Civic networks are embodiments of distinctive technological configurations and forms of social order. In choosing some technological configurations over others, designers are favoring some social structural configurations over alternatives. To the extent that a civic network sets out to reconfigure the prevailing social order (as was the case in the Urban-net project considered here), the design process becomes the arena where challengers of the prevailing order encounter its defenders. In this case the defenders prevailed, and the design that emerged was conservative and reproduced the status quo. What steps can stakeholders take so that the project’s future development is in line with the original aim of structural change? We outline two strategies. We argue the importance of articulating cultural desiderata in an architecture that stakeholders can use to open up the infrastructure to new constituents and incremental change. Next, we argue the importance of designing the conditions of design. The climate in which social interactions occur can powerfully shape design outcomes, but this does not usually figure in stakeholders’ design concerns.     

Adaptivity in hybrid-Trefftz finite element formulation

With the advent of computer aided design (CAD), the development of fully or semi-automated procedures by which a solution warranting the specified accuracy can most efficiently be reached has become a necessity. It is now of utmost interest to investigate finite element (FE) formulations offering a simple form of adaptivity, easy to implement. The present study (which is a sequel to an earlier paper¹ on error estimation by the authors) identifies as a particularly promising approach of attaining this goal the p-method associated with the so-called hybrid-Trefftz FE model.² Based on a simple stress error estimator¹ and prior knowledge of the convergence rate, the solution warranting specified stress accuracy may be reached in a single trial by suitably respecifying just one parameter in the input data. The reported approach has successfully been implemented into the general purpose FE program SAFE³ and its high efficiency is illustrated in the paper through practical applications involving corner singularities and stress concentrations.     

An optimal control analysis for baseflow externalities in an interconnected lake–aquifer system

Groundwater resources play a vital role in regional growth and development. From a sustainability viewpoint, it is, however, imperative to ensure that groundwater extractions do not lead to deleterious environmental consequences. The interconnections between groundwater in the aquifers and surface water bodies are increasingly being recognized in recent times and groundwater discharges to surface water bodies are critical to maintain water levels and water quality during summer months and periods of drought. Excessive groundwater withdrawals can cause water levels in streams and lakes to drop below unacceptable levels. This phenomenon is referred to as baseflow externality in this study. The primary goal of this investigation was to develop a mathematical model useful to assess policy options pertaining to groundwater withdrawals considering baseflow externalities. The model was used to evaluate five wide-ranging policy options that emphasize economic gains and ecological impacts to different levels. The weak sustainability policy option represents a reasonable compromise and identifies groundwater pumping that maximizes economic benefits while maintaining acceptable water levels in an interconnected lake. However, the implementation success of this approach hinges on achieving consensus among stakeholders and the ability to find reliable additional water supplies at the end of the planning period. The developed model can be a useful tool for facilitating stakeholder discussions that foster participatory groundwater management.     

Recent developments in grinding of advanced materials

This article discusses the recent developments in grinding of advanced materials. Eighty-four journal papers published recently are briefly introduced. The topics are advances in grinding of brittle materials, grinding of silicon, dressing/truing of grinding wheels, grinding fluids, grinding of mirrors and vibration-assisted grinding, measuring/monitoring of grinding, optimization of grinding, modelling and simulation of grinding, and size effect. Ductile mode grinding of brittle materials has been and will continue to be an intensive research area because of its increasing industrial applications and academic demands for fundamental understanding of the ductile mode grinding mechanism. Highly precision manufacturing of silicon substrates faces more and more new challenges. Grinding of silicon continues to be a popular research topic. Using lasers to true and dress grinding wheels has attracted great research interest, because it has significant advantages over mechanical processes. Environmentally friendly grinding fluids are increasingly highly demanded. Vibration-assisted grinding is promising. Monitoring, modelling and optimization of grinding processes help to understand grinding mechanisms and achieve better grinding performance. The size effect is more prominent in grinding than turning and can be used for obtaining a controlled work-hardening surface layer with higher wear resistance and hardness.     

Rapid estimation and quantification of sucrose content in fruit juices using Fourier transform infrared–attenuated total reflectance (FTIR–ATR) spectroscopy

An uncomplicated and rapid procedure has been developed for the quantitative analysis of sucrose in fruit samples (grape, pineapple, mango) through attenuated total reflectance–Fourier transform infrared absorbance measurements (ATR–FTIR). FTIR analysis takes considerably reduced time compared to the other classical methods. To calibrate the method, we used firstly, different concentrations of pure sucrose (from 1 to 5 %) and registered their IR maximal wavenumbers and peak intensity. The spectral peak of sucrose for each sample lies between 1057 and 1061 cm^?1. DNS method was used to analyse the content of sucrose by using spectrophotometry. The wave length used for analysing is 540 nm. Also high performance liquid chromatography was used to analyse the sucrose content in the fruit juices. By comparing the retention time of sucrose standards and the sample juices, sucrose concentration was identified and quantified. The results of all three experiments/techniques support each other by justifying that the mango has the high content of sucrose followed by pineapple and grape.