Abstract
Apresentamos o resultado de uma exploração prática da tecnologia de mundos virtuais imersivos multiutilizador High Fidelity, baseada em tecnologia Web. Esta tecnologia permite criar mundos virtuais cujas formas de interligação, controlo e interação tiram partido de uma Interface de Programação de Aplicações em JavaScript. Através do caso prático de desenvolvimento de um protótipo de jogo educativo simples, descrevemos a tecnologia High Fidelity, incluindo o tipo de scripts, a arquitetura inerente e as suas características de desenvolvimento e utilização. Destacamos as dificuldades inerentes ao estado atual da plataforma, em constante reformulação e algumas peculiaridades.

Abstract
Cardiotocography is a diagnostic exam performed from the 28th week of pregnancy that registers the fetus cardiac frequency and uterine contractions. From this exam results a cardiotocogram whose reading and observation of the patterns contained in it allow an evaluation of the baby's condition and the fetal vitality in the maternal womb. This work aims the creation of a classification model using Learning Algorithms/Data Mining using the tool Rapid Miner. The subject of study was a Data Set with information registered from a total of 2126 cardiotograms, with 23 attributes, properly classified by 3 specialized obstetricians as to the baby status, in three possible states, namely: N = Normal; S = Suspect; P = Pathologic. All models tested showed an overall accuracy greater than 80%. Therefore the usefulness of creating predictive models for the classification of this type of diagnosis is great.

Abstract
Stochastic search algorithms are black-box optimizer of an objective function. They have recently gained a lot of attention in operations research, machine learning and policy search of robot motor skills due to their ease of use and their generality. Yet, many stochastic search algorithms require relearning if the task or objective function changes slightly to adapt the solution to the new situation or the new context. In this paper, we consider the contextual stochastic search setup. Here, we want to find multiple good parameter vectors for multiple related tasks, where each task is described by a continuous context vector. Hence, the objective function might change slightly for each parameter vector evaluation of a task or context. Contextual algorithms have been investigated in the field of policy search, however, the search distribution typically uses a parametric model that is linear in the some hand-defined context features. Finding good context features is a challenging task, and hence, non-parametric methods are often preferred over their parametric counter-parts. In this paper, we propose a non-parametric contextual stochastic search algorithm that can learn a non-parametric search distribution for multiple tasks simultaneously. In difference to existing methods, our method can also learn a context dependent covariance matrix that guides the exploration of the search process. We illustrate its performance on several non-linear contextual tasks.

Abstract
Aims: Cardiovascular (CV) events are the leading cause of morbidity and mortality in patients with chronic kidney disease (CKD), including those patients on peritoneal dialysis (PD). Fibroblast growth factor 23 (FGF23) has been associated with left ventricular hypertrophy (LVH) and mortality in patients with CKD. However, the role of FGF23 in uremic vasculopathy remains unclear. In this study, we aimed to assess the relationship between FGF23 and LVH, endothelial dysfunction, vascular calcification, and arterial stiffness in 48 stable PD patients. Methods: Left ventricular mass index (LVMI) was assessed using 2-D echocardiography. Intact FGF23 blood levels were evaluated using an ELISA kit (Immutopics, Inc., San Clemente, CA, USA). Reactive hyperemia index (RHI) is a surrogate marker of endothelial dysfunction and the augmentation index (AI) is a surrogate marker of arterial stiffness. Both were assessed using peripheral arterial tonometry (EndoPAT 2000). Vascular calcification (VC) was assessed using the Adragao score. Results: In unadjusted analysis; FGF23 was positively correlated with serum Pi (r = 0.487, p < 0.001), serum urea (r = 0.351, p = 0.015), serum creatinine (r = 0.535, p < 0.001), dialysis vintage (r = 0.309, p = 0.033), and LVMI (r = 0.369, p = 0.027) and was negatively correlated with age (r = -0.343, p = 0.017), residual renal function (r = -0.359, p < 0.012), and AI (r = -0.304, p = 0.038). In multivariate adjusted analysis, FGF23 was associated with LVMI (beta = 0.298, p = 0.041), serum Pi (beta = 0.345, p = 0.018), and age (beta = -0.372, p = 0.007) independent of dialysis vintage, gender, residual renal function (RRF), albumin, C-reactive protein and systolic blood pressure. There were no associations found between FGF23 and RHI, AI, or VC in multivariable-adjusted models. Conclusions: Our results show that FGF23 is associated with LVH but not with endothelial dysfunction, arterial stiffness, or vascular calcification in PD patients.

Abstract
Multiyear Transmission Expansion Planning (TEP) aims to determine how and when a transmission network capacity should be expanded taking into account an extended horizon. This is an optimization problem very difficult to solve and that has unique characteristics that increase its complexity such as its non-convex search space and its integer and nonlinear nature. This paper describes a hybrid tool to solve the TEP problem, including a first phase to select a list of equipment candidates conducted by a Constructive Heuristic Algorithm (CHA), and a second phase that uses Discrete Evolutionary Particle Swarm Optimization (DEPSO) for the final planning. Both phases use the AC power flow model as a way to improve the realism of the developed tool. The paper includes a case study based on the IEEE 24-Bus Reliability Test System and the results show that tools based on swarm intelligence applied to reduced search spaces are able to find good quality solutions with low computational effort.

Abstract
In this paper a novel scenario generation methodology based on artificial neural networks (ANNs) is proposed. The methodology is flexible and able to generate scenarios for various stochastic variables that are used as input parameters in the stochastic short-term scheduling models. Appropriate techniques for modeling the cross-correlation of the involved stochastic processes and scenario reduction techniques are also incorporated into the proposed approach. The applicability of the methodology is investigated through the creation of electric load, photovoltaic (PV) and wind production scenarios and the performance of the proposed ANN-based methodology is compared to time series-based scenario generation models. Test results on the real-world insular power system of Crete and mainland Greece present the effectiveness of the proposed ANN-based scenario generation methodology.