A sample covering 204,394 blocks from the 1990 U.S. Census permits measurement of residual heterogeneity from local area to local area after controlling by stratification for demographic characteristics such as race, ethnicity, age, sex as well as geographic characteristics such as region and place-type. The local areas have populations on the order of 10,000 people. The variables studied are f...

We investigate a local reparameterizaton technique for greatly reducing the variance of stochastic gradients for variational Bayesian inference (SGVB) of a posterior over model parameters, while retaining parallelizability. This local reparameterization translates uncertainty about global parameters into local noise that is independent across datapoints in the minibatch. Such parameterizations ...

We study the stochastic Riemannian gradient algorithm for matrix eigendecomposition. The state-of-the-art stochastic Riemannian algorithm requires the learning rate to decay to zero and thus suffers from slow convergence and suboptimal solutions. In this paper, we address this issue by deploying the variance reduction (VR) technique of stochastic gradient descent (SGD). The technique was origin...

Where observers concentrate their gaze during visual search depends on several factors. The aim here was to determine how much of the variance in observers' fixations in natural scenes can be explained by local scene color and how that variance is related to viewing bias. Fixation data were taken from an experiment in which observers searched images of 20 natural rural and urban scenes for a sm...

The current study reports upon further evaluation of the psychometric properties of the Learning and Development Survey. Originally tested on 113 local government employees, the Learning and Development Survey was found to have good reliability, however there was a large proportion of redundant items, with 62 indicators spanning six constructs. These six constructs were: Organisational Opportun...

Masry (1996b) provides estimation bias and variance expression for a general local polynomial kernel estimator in a general multivariate regression framework. Under smoother conditions on the unknown regression and by including more refined approximation terms than that in Masry (1996b), we extend the result of Masry (1996b) to obtain explicit leading bias terms for the whole vector of the loca...

This paper studies the properties of the quasi -maximum likelihood estimator (QMLE) and related test statistics in dynamic models that jointly parameterize conditional means and conditional covariances when a normal log likelihood is maximized but the assumption of normality is violated. Because the score of the normal log likelihood has the martingale difference property under fairly general r...

We report on the adequacy of using Sato processes to value equity structured products. In models used to price options on realized variance, the latter must be a random variable with a positive variance. An analysis of this variance of realized variance for Sato processes shows that these processes may be suited to option contracts on realized volatility. Nonlinear pricing principles based on h...

Managing risk in dynamic decision problems is of cardinal importance in many fields such as finance and process control. The most common approach to defining risk is through various variance related criteria such as the Sharpe Ratio or the standard deviation adjusted reward. It is known that optimizing many of the variance related risk criteria is NP-hard. In this paper we devise a framework fo...