Gaussian processes:iterative sparse approximations

In recent years there has been an increased interest in applying non-parametric methods to real-world problems. Significant research has been devoted to Gaussian processes (GPs) due to their increased flexibility when compared with parametric models. These methods use Bayesian learning, which generally leads to analytically intractable posteriors. This thesis proposes a two-step solution to construct a probabilistic approximation to the posterior. In the first step we adapt the Bayesian online learning to GPs: the final approximation to the posterior is the result of propagating the first and second moments of intermediate posteriors obtained by combining a new example with the previous approximation. The propagation of em functional forms is solved by showing the existence of a parametrisation to posterior moments that uses combinations of the kernel function at the training points, transforming the Bayesian online learning of functions into a parametric formulation. The drawback is the prohibitive quadratic scaling of the number of parameters with the size of the data, making the method inapplicable to large datasets. The second step solves the problem of the exploding parameter size and makes GPs applicable to arbitrarily large datasets. The approximation is based on a measure of distance between two GPs, the KL-divergence between GPs. This second approximation is with a constrained GP in which only a small subset of the whole training dataset is used to represent the GP. This subset is called the em Basis Vector, or BV set and the resulting GP is a sparse approximation to the true posterior. As this sparsity is based on the KL-minimisation, it is probabilistic and independent of the way the posterior approximation from the first step is obtained. We combine the sparse approximation with an extension to the Bayesian online algorithm that allows multiple iterations for each input and thus approximating a batch solution. The resulting sparse learning algorithm is a generic one: for different problems we only change the likelihood. The algorithm is applied to a variety of problems and we examine its performance both on more classical regression and classification tasks and to the data-assimilation and a simple density estimation problems.

Femtosecond laser micro-inscription of optical coherence tomography resolution test artifacts

Optical coherence tomography (OCT) systems are becoming more commonly used in biomedical imaging and, to enable continued uptake, a reliable method of characterizing their performance and validating their operation is required. This paper outlines the use of femtosecond laser subsurface micro-inscription techniques to fabricate an OCT test artifact for validating the resolution performance of a commercial OCT system. The key advantage of this approach is that by utilizing the nonlinear absorption a three dimensional grid of highly localized point and line defects can be written in clear fused silica substrates.

Alogliptin after acute coronary syndrome in patients with type 2 diabetes

Background - To assess potentially elevated cardiovascular risk related to new antihyperglycemic drugs in patients with type 2 diabetes, regulatory agencies require a comprehensive evaluation of the cardiovascular safety profile of new antidiabetic therapies. We assessed cardiovascular outcomes with alogliptin, a new inhibitor of dipeptidyl peptidase 4 (DPP-4), as compared with placebo in patients with type 2 diabetes who had had a recent acute coronary syndrome. Methods - We randomly assigned patients with type 2 diabetes and either an acute myocardial infarction or unstable angina requiring hospitalization within the previous 15 to 90 days to receive alogliptin or placebo in addition to existing antihyperglycemic and cardiovascular drug therapy. The study design was a double-blind, noninferiority trial with a prespecified noninferiority margin of 1.3 for the hazard ratio for the primary end point of a composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. Results - A total of 5380 patients underwent randomization and were followed for up to 40 months (median, 18 months). A primary end-point event occurred in 305 patients assigned to alogliptin (11.3%) and in 316 patients assigned to placebo (11.8%) (hazard ratio, 0.96; upper boundary of the one-sided repeated confidence interval, 1.16; P<0.001 for noninferiority). Glycated hemoglobin levels were significantly lower with alogliptin than with placebo (mean difference, -0.36 percentage points; P<0.001). Incidences of hypoglycemia, cancer, pancreatitis, and initiation of dialysis were similar with alogliptin and placebo. Conclusions - Among patients with type 2 diabetes who had had a recent acute coronary syndrome, the rates of major adverse cardiovascular events were not increased with the DPP-4 inhibitor alogliptin as compared with placebo. (Funded by Takeda Development Center Americas; EXAMINE ClinicalTrials.gov number, NCT00968708.)

E-HR and international HRM:a critical perspective on the discursive framing of e-HR

In the current global economic climate, international HRM is facing unprecedented pressure to become more innovative, effective and efficient. New discourses are emerging around the application of information technology, with 'e-HR' (electronic-enablement of Human Resources), self-service portals and promises of improved services couched as various HR 'value propositions'. This study explores these issues through our engagement with the emergent stream of 'critical' HRM, the broader study of organizational discourse and ethical management theories. We have found that while there is growing research into the take-up of e-HR applications, there is a dearth of investigation into the impact of e-HR on the people involved; in particular, the (re)structuring of social relations between HR functions and line managers in the move away from face-to-face HR support services, to more technology-mediated 'self-service' relationships. We undertake a close reading of personal narratives from a multinational organization, deploying a critical discourse lens to examine different dimensions of e-HR and raise questions about the strong technocratic framing of the international language of people management, shaping line manager enactment of e-HR duties. We argue for a more reflexive stance in the conceptualization e-HR, and conclude with a discussion about the theoretical and practical implications of our study, limitations and suggestions for future research. © 2014 © 2014 Taylor & Francis.

A variational conjugate gradient method for determining the fluid velocity of a slow viscous flow

The problem considered is that of determining the fluid velocity for linear hydrostatics Stokes flow of slow viscous fluids from measured velocity and fluid stress force on a part of the boundary of a bounded domain. A variational conjugate gradient iterative procedure is proposed based on solving a series of mixed well-posed boundary value problems for the Stokes operator and its adjoint. In order to stabilize the Cauchy problem, the iterations are ceased according to an optimal order discrepancy principle stopping criterion. Numerical results obtained using the boundary element method confirm that the procedure produces a convergent and stable numerical solution.