A simple iterative algorithm for parsimonious binary kernel Fisher discrimination

By applying recent results in optimisation theory variously known as optimisation transfer or majorise/minimise algorithms, an algorithm for binary, kernel, Fisher discriminant analysis is introduced that makes use of a non-smooth penalty on the coefficients to provide a parsimonious solution. The problem is converted into a smooth optimisation that can be solved iteratively with no greater overhead than iteratively re-weighted least-squares. The result is simple, easily programmed and is shown to perform, in terms of both accuracy and parsimony, as well as or better than a number of leading machine learning algorithms on two well-studied and substantial benchmarks.     

Nonparabolicity effects and the spin-split electron dwell time in symmetric III-V double-barrier structures

We start from the fourth-order nonparabolic and anisotropic conduction band bulk dispersion relation to obtain a one-band effective Hamiltonian, which we apply to an AlGaSb symmetric double-barrier structure with resonant energies significantly (more than 200 meV) above the well bottom. The spin-splitting is described by the k³ Dresselhaus spin-orbit coupling term modifying only the effective mass of the spin eigenstates in the investigated structure. Apart from the bulk-like resonant energy shift due to the band nonparabolicity, we obtain a substantial shift depending on the choice of boundary conditions for the envelope functions at interfaces between different materials. The shift of resonant energy levels leads to the change of spin-splitting and the magnitude of the dwell times. We attempt to explain the influence of both the nonparabolicity and boundary conditions choice by introducing various effective masses.     

Setting performance criteria for building products

As previously reported in our pages, the CIB working group on the performance concept in building, W60, has produced a booklet of six papers, each dealing with setting performance criteria for building products. They present alternative methodologies and are not intended to provide detailed guidance for the values to be specified. The document is summarised and some of the major issues involved are outlined in this article. There is only one fundamental criterion of success, it is concluded, namely the correct matching of the solution to the originally-defined user needs. The different methods discussed are all means to that end.     

Experimental Characterization and Optimization of Hybrid FRP/RC Bridge Superstructure System

An experimental investigation was performed to assess the performance of a hybrid fiber-reinforced polymer/reinforced concrete bridge system. The full-scale laboratory specimen was representative of an 813?mm (32?in.) wide strip of a completed bridge in San Patricio County, Tex. The specimen was first subjected to static loading prior to casting the reinforced concrete deck. Displacement, strain, and acoustic emission were recorded. After completion of the nondestructive static loading a reinforced concrete deck was cast in the laboratory to represent one unit of the completed bridge. Load was statically applied with several increased load cycles until failure occurred at a load level exceeding 18 times the calculated design load. The results of the static testing indicated that the original design of the hybrid bridge was very conservative. An optimized design of the hybrid bridge was then derived. The static load testing program and the resulting optimized design are described.     

SnGe Asymmetric Quantum Well Electroabsorption Modulators for Long-Wave Silicon Photonics

In recent years, SiGeSn alloys have been considered as promising materials for optoelectronic applications because they offer the possibility of a direct bandgap, and are compatible with Si-based technology; therefore, they have prospective applications such as in interband lasers and detectors, solar cells, etc. Here, we consider another possible application of nanostructures based on these materials: to extend the suite of Si-based optoelectronic devices for interband electroabsorption (EA) modulators. Using the 8-band $hbox{bf k.p}$ method, we have designed asymmetric double quantum wells that are optimized, by varying the well and barrier widths and material composition, to show large optical transmission sensitivity to the applied bias. Generally, these structures are useful for EA modulators in the midinfrared spectral range.      

Understanding attitudes toward affirmative action programs in employment: Summary and meta-analysis of 35 years of research.

Affirmative action programs (AAPs) are controversial employment policies in the United States and elsewhere. A large body of evidence about attitudinal reactions to AAPs in employment has accumulated over 35 years: at least 126 independent samples involving 29,000 people. However, findings are not firmly established or integrated. In the current article, the authors summarize and meta-analytically estimate relationships of AAP attitudes with (a) structural features of such programs, (b) perceiver demographic and psychological characteristics, (c) interactions of structural features with perceiver characteristics, and (d) presentation of AAP details to perceivers, including justification of the AAP. Results are generally consistent with predictions derived from self-interest considerations, organizational justice theory, and racism theories. They also suggest practical ways in which AAPs might be designed and communicated to employees to reduce attitudinal resistance. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Design and implementation of multipattern generators in analog VLSI

In recent years, computational biologists have shown through simulation that small neural networks with fixed connectivity are capable of producing multiple output rhythms in response to transient inputs. It is believed that such networks may play a key role in certain biological behaviors such as dynamic gait control. In this paper, we present a novel method for designing continuous-time recurrent neural networks (CTRNNs) that contain multiple embedded limit cycles, and we show that it is possible to switch the networks between these embedded limit cycles with simple transient inputs. We also describe the design and testing of a fully integrated four-neuron CTRNN chip that is used to implement the neural network pattern generators. We provide two example multipattern generators and show that the measured waveforms from the chip agree well with numerical simulations.     

Links between environment and health: possible future directions

Linkages between environmental quality and human health are likely to continue to be of major interest, but quantitatively smaller risks impacting on larger population groups will probably become a prime future focus. Increasingly the aim will be to understand the inter-relatedness of impacts on human health and to sustain human physical and mental well-being rather than simply preventing disease.     

Trapping of bead-based reagents within microfluidic systems: on-chip solid-phase extraction and electrochromatography

A 330-pL chromatographic bed was fabricated on a glass substrate as part of an electroosmotically pumped microfluidic system. Two weirs within a sample channel formed a cavity in which octadecylsilane (ODS) coated silica beads (1.5-4 microns diameter) were trapped. ODS beads were mobilized into and out of the cavity using electroosmotic flow through a bead-introduction channel which accessed the cavity. This procedure allowed the beads in the cavity to be repeatedly exchanged. A 1 nM solution of a nonpolar analyte (BODIPY 493/503) in buffer was loaded onto the beads for different lengths of time using an electroosmotic flow of 1.2 nL/s. The material retained on the ODS phase was then eluted by electroosmotic flow of acetonitrile with a concentration enhancement of up to 500 times. Capillary electrochromatography was conducted using a similar device. BODIPY and fluorescein were loaded onto a 200-micron-long chromatographic bed and then separated in an isocratic CEC run with an acetonitrile/buffer mobile phase. Complete separation was achieved in less than 20 s with a 2-micron plate height.     

Nutritive value of corn silage as affected by maturity and mechanical processing: a contemporary review

Stage of maturity at harvest and mechanical processing affect the nutritive value of corn silage. The change in nutritive value of corn silage as maturity advances can be measured by animal digestion and macro in situ degradation studies among other methods. Predictive equations using climatic data, vitreousness of corn grain in corn silage, starch reactivity, gelatinization enthalpy, dry matter (DM) of corn grain in corn silage, and DM of corn silage can be used to estimate starch digestibility of corn silage. Whole plant corn silage can be mechanically processed either pre- or postensiling with a kernel processor mounted on a forage harvester, a recutter screen on a forage harvester, or a stationary roller mill. Mechanical processing of corn silage can improve ensiling characteristics, reduce DM losses during ensiling, and improve starch and fiber digestion as a result of fracturing the corn kernels and crushing and shearing the stover and cobs. Improvements in milk production have ranged from 0.2 to 2.0 kg/d when cows were fed mechanically processed corn silage. A consistent improvement in milk protein yield has also been observed when mechanically processed corn silage has been fed. With the advent of mechanical processors, alternative strategies are evident for corn silage management, such as a longer harvest window.