Error and variance bounds on sigmoidal neurons with weight and input errors

Bounds on the expectation and variance of errors at the output of a multilayer feedforward neural network with perturbed weights and inputs are derived. It is assumed that errors in weights and inputs to the network are statistically independent and small. The bounds obtained are applicable to both digital and analogue network implementations and are shown to be of practical value.

A method to quantify a descriptor's illumination variance

This paper presents a new metric, which we call the lighting variance ratio, for quantifying descriptors in terms of their variance to illumination changes. In many applications it is desirable to have descriptors that are robust to changes in illumination, especially in outdoor environments. The lighting variance ratio is useful for comparing descriptors and determining if a descriptor is lighting invariant enough for a given environment. The metric is analysed across a number of datasets, cameras and descriptors. The results show that the upright SIFT descriptor is typically the most lighting invariant descriptor.

Multivariate variance-components analysis in DTI

Twin studies are a major research direction in imaging genetics, a new field, which combines algorithms from quantitative genetics and neuroimaging to assess genetic effects on the brain. In twin imaging studies, it is common to estimate the intraclass correlation (ICC), which measures the resemblance between twin pairs for a given phenotype. In this paper, we extend the commonly used Pearson correlation to a more appropriate definition, which uses restricted maximum likelihood methods (REML). We computed proportion of phenotypic variance due to additive (A) genetic factors, common (C) and unique (E) environmental factors using a new definition of the variance components in the diffusion tensor-valued signals. We applied our analysis to a dataset of Diffusion Tensor Images (DTI) from 25 identical and 25 fraternal twin pairs. Differences between the REML and Pearson estimators were plotted for different sample sizes, showing that the REML approach avoids severe biases when samples are smaller. Measures of genetic effects were computed for scalar and multivariate diffusion tensor derived measures including the geodesic anisotropy (tGA) and the full diffusion tensors (DT), revealing voxel-wise genetic contributions to brain fiber microstructure.

Analysis of differential gene expression based on Bayesian estimation of variance

Gene expression is arguably the most important indicator of biological function. Thus identifying differentially expressed genes is one of the main aims of high throughout studies that use microarray and RNAseq platforms to study deregulated cellular pathways. There are many tools for analysing differentia gene expression from transciptomic datasets. The major challenge of this topic is to estimate gene expression variance due to the high amount of ‘background noise’ that is generated from biological equipment and the lack of biological replicates. Bayesian inference has been widely used in the bioinformatics field. In this work, we reveal that the prior knowledge employed in the Bayesian framework also helps to improve the accuracy of differential gene expression analysis when using a small number of replicates. We have developed a differential analysis tool that uses Bayesian estimation of the variance of gene expression for use with small numbers of biological replicates. Our method is more consistent when compared to the widely used cyber-t tool that successfully introduced the Bayesian framework to differential analysis. We also provide a user-friendly web based Graphic User Interface for biologists to use with microarray and RNAseq data. Bayesian inference can compensate for the instability of variance caused when using a small number of biological replicates by using pseudo replicates as prior knowledge. We also show that our new strategy to select pseudo replicates will improve the performance of the analysis. - See more at: http://www.eurekaselect.com/node/138761/article#sthash.VeK9xl5k.dpuf

Enhanced sensitivity in detection of Kerr rotation by double modulation and time averaging based on Allan variance

Experiments in spintronics necessarily involve the detection of spin polarization. The sensitivity of this detection becomes an important factor to consider when extending the low temperature studies on semiconductor spintronic devices to room temperature, where the spin signal is weaker. In pump-probe experiments, which optically inject and detect spins, the sensitivity is often improved by using a photoelastic modulator (PEM) for lock-in detection. However, spurious signals can arise if diode lasers are used as optical sources in such experiments, along with a PEM. In this work, we eliminated the spurious electromagnetic coupling of the PEM onto the probe diode laser, by the double modulation technique. We also developed a test for spurious modulated interference in the pump-probe signal, due to the PEM. Besides, an order of magnitude enhancement in the sensitivity of detection of spin polarization by Kerr rotation, to 3x10(-8) rad was obtained by using the concept of Allan variance to optimally average the time series data over a period of 416 s. With these improvements, we are able to experimentally demonstrate at room temperature, photoinduced steady-state spin polarization in bulk GaAs. Thus, the advances reported here facilitate the use of diode lasers with a PEM for sensitive pump-probe experiments. They also constitute a step toward detection of spin-injection in Si at room temperature.

Modeling strategies in longitudinal data analysis: Covariate, variance function and correlation structure selection

A modeling paradigm is proposed for covariate, variance and working correlation structure selection for longitudinal data analysis. Appropriate selection of covariates is pertinent to correct variance modeling and selecting the appropriate covariates and variance function is vital to correlation structure selection. This leads to a stepwise model selection procedure that deploys a combination of different model selection criteria. Although these criteria find a common theoretical root based on approximating the Kullback-Leibler distance, they are designed to address different aspects of model selection and have different merits and limitations. For example, the extended quasi-likelihood information criterion (EQIC) with a covariance penalty performs well for covariate selection even when the working variance function is misspecified, but EQIC contains little information on correlation structures. The proposed model selection strategies are outlined and a Monte Carlo assessment of their finite sample properties is reported. Two longitudinal studies are used for illustration.

Effects of variance-function misspecification in analysis of longitudinal data

The approach of generalized estimating equations (GEE) is based on the framework of generalized linear models but allows for specification of a working matrix for modeling within-subject correlations. The variance is often assumed to be a known function of the mean. This article investigates the impacts of misspecifying the variance function on estimators of the mean parameters for quantitative responses. Our numerical studies indicate that (1) correct specification of the variance function can improve the estimation efficiency even if the correlation structure is misspecified; (2) misspecification of the variance function impacts much more on estimators for within-cluster covariates than for cluster-level covariates; and (3) if the variance function is misspecified, correct choice of the correlation structure may not necessarily improve estimation efficiency. We illustrate impacts of different variance functions using a real data set from cow growth.

Brownian particles in stationary and moving traps: The mean and variance of the heat distribution function

A recent theoretical model developed by Imparato et al. Phys of the experimentally measured heat and work effects produced by the thermal fluctuations of single micron-sized polystyrene beads in stationary and moving optical traps has proved to be quite successful in rationalizing the observed experimental data. The model, based on the overdamped Brownian dynamics of a particle in a harmonic potential that moves at a constant speed under a time-dependent force, is used to obtain an approximate expression for the distribution of the heat dissipated by the particle at long times. In this paper, we generalize the above model to consider particle dynamics in the presence of colored noise, without passing to the overdamped limit, as a way of modeling experimental situations in which the fluctuations of the medium exhibit long-lived temporal correlations, of the kind characteristic of polymeric solutions, for instance, or of similar viscoelastic fluids. Although we have not been able to find an expression for the heat distribution itself, we do obtain exact expressions for its mean and variance, both for the static and for the moving trap cases. These moments are valid for arbitrary times and they also hold in the inertial regime, but they reduce exactly to the results of Imparato et al. in appropriate limits. DOI: 10.1103/PhysRevE.80.011118 PACS.

Self-tuning minimum-variance control of nonlinear systems of the Hammerstein model

Self-tuning is applied to the control of nonlinear systems represented by the Hammerstein model wherein the nonlinearity is any odd-order polynomial. But control costing is not feasible in general. Initial relay control is employed to contain the deviations.

No variance for filed orders: Hayes v Westpac Banking Corporation [2015] QCA 260

In Hayes v Westpac Banking Corporation [2015] QCA 260 the Queensland Court of Appeal examined the relationship between rules 7 (extending and shortening time) and 667 (setting aside) of the Uniform Civil Procedure Rules 1999 (Qld), and held that r667(1) does not enable the court to set aside or vary an order after the order has been filed. The court found that, to the extent that this conclusion was contrary to the decision in McIntosh v Linke Nominees Pty Ltd [2010] 1 Qd R 152, the decision in McIntosh was wrong and should not be followed.

Evolution of polyandry by reduction in progeny number variance in structured populations

When there is a variation in the quality of males in a population, multiple mating can lead to an increase in the genetic fitness of a female by reducing the variance of the progeny number. The extent of selective advantage obtainable by this process is investigated for a population subdivided into structured demes. It is seen that for a wide range of model parameters (deme size, distribution of male quality, local resource level), multiple mating leads to a considerable increase in the fitness. Frequency-dependent selection or a stable coexistence between polyandry and monandry can also result when the possible costs involved in multiple mating are taken into account.

Modelling and analysis of the variance in parallelism in parallel computations

In this paper, we introduce an analytical technique based on queueing networks and Petri nets for making a performance analysis of dataflow computations when executed on the Manchester machine. This technique is also applicable for the analysis of parallel computations on multiprocessors. We characterize the parallelism in dataflow computations through a four-parameter characterization, namely, the minimum parallelism, the maximum parallelism, the average parallelism and the variance in parallelism. We observe through detailed investigation of our analytical models that the average parallelism is a good characterization of the dataflow computations only as long as the variance in parallelism is small. However, significant difference in performance measures will result when the variance in parallelism is comparable to or higher than the average parallelism.

A Linear S-N Curve with Load Dependent Variance and Explicit Failure Probability

We reconsider standard uniaxial fatigue test data obtained from handbooks. Many S-N curve fits to such data represent the median life and exclude load-dependent variance in life. Presently available approaches for incorporating probabilistic aspects explicitly within the S-N curves have some shortcomings, which we discuss. We propose a new linear S-N fit with a prespecified failure probability, load-dependent variance, and reasonable behavior at extreme loads. We fit our parameters using maximum likelihood, show the reasonableness of the fit using Q-Q plots, and obtain standard error estimates via Monte Carlo simulations. The proposed fitting method may be used for obtaining S-N curves from the same data as already available, with the same mathematical form, but in cases in which the failure probability is smaller, say, 10 % instead of 50 %, and in which the fitted line is not parallel to the 50 % (median) line.