Desarrollo de software para el procesado numérico en tarjetas gráficas

Debido al gran número de transistores por mm2 que hoy en día podemos encontrar en las GPU convencionales, en los últimos años éstas se vienen utilizando para propósitos generales gracias a que ofrecen un mayor rendimiento para computación paralela. Este proyecto implementa el producto sparse matrix-vector sobre OpenCL. En los primeros capítulos hacemos una revisión de la base teórica necesaria para comprender el problema. Después veremos los fundamentos de OpenCL y del hardware sobre el que se ejecutarán las librerías desarrolladas. En el siguiente capítulo seguiremos con una descripción del código de los kernels y de su flujo de datos. Finalmente, el software es evaluado basándose en comparativas con la CPU.

An empirical analysis of a merger between a network and low-cost airlines

We use a difference-in-difference estimator to examine the effects of a merger involving three airlines. The novelty lies in the examination of this operation in two distinct scenarios: (1) on routes where two low-cost carriers and (2) on routes where a network and one of the low-cost airlines had previously been competing. We report a reduction in frequencies but no substantial effect on prices in the first scenario, while in the second we report an increase in prices but no substantial effect on frequencies. These results may be attributed to the differences in passenger types flying on these routes.

Estimation of parametric and nonparametric models for univariate claim severity distributions - an approach using R

This paper presents an analysis of motor vehicle insurance claims relating to vehicle damage and to associated medical expenses. We use univariate severity distributions estimated with parametric and non-parametric methods. The methods are implemented using the statistical package R. Parametric analysis is limited to estimation of normal and lognormal distributions for each of the two claim types. The nonparametric analysis presented involves kernel density estimation. We illustrate the benefits of applying transformations to data prior to employing kernel based methods. We use a log-transformation and an optimal transformation amongst a class of transformations that produces symmetry in the data. The central aim of this paper is to provide educators with material that can be used in the classroom to teach statistical estimation methods, goodness of fit analysis and importantly statistical computing in the context of insurance and risk management. To this end, we have included in the Appendix of this paper all the R code that has been used in the analysis so that readers, both students and educators, can fully explore the techniques described

How does monetary policy change? evidence on inflation targeting countries

We examine the evolution of monetary policy rules in a group of inflation targeting countries (Australia, Canada, New Zealand, Sweden and the United Kingdom) applying moment- based estimator at time-varying parameter model with endogenous regressors. Using this novel flexible framework, our main findings are threefold. First, monetary policy rules change gradually pointing to the importance of applying time-varying estimation framework. Second, the interest rate smoothing parameter is much lower that what previous time-invariant estimates of policy rules typically report. External factors matter for all countries, albeit the importance of exchange rate diminishes after the adoption of inflation targeting. Third, the response of interest rates on inflation is particularly strong during the periods, when central bankers want to break the record of high inflation such as in the U.K. or in Australia at the beginning of 1980s. Contrary to common wisdom, the response becomes less aggressive after the adoption of inflation targeting suggesting the positive effect of this regime on anchoring inflation expectations. This result is supported by our finding that inflation persistence as well as policy neutral rate typically decreased after the adoption of inflation targeting.

Indirect likelihood inference

Given a sample from a fully specified parametric model, let Zn be a given finite-dimensional statistic - for example, an initial estimator or a set of sample moments. We propose to (re-)estimate the parameters of the model by maximizing the likelihood of Zn. We call this the maximum indirect likelihood (MIL) estimator. We also propose a computationally tractable Bayesian version of the estimator which we refer to as a Bayesian Indirect Likelihood (BIL) estimator. In most cases, the density of the statistic will be of unknown form, and we develop simulated versions of the MIL and BIL estimators. We show that the indirect likelihood estimators are consistent and asymptotically normally distributed, with the same asymptotic variance as that of the corresponding efficient two-step GMM estimator based on the same statistic. However, our likelihood-based estimators, by taking into account the full finite-sample distribution of the statistic, are higher order efficient relative to GMM-type estimators. Furthermore, in many cases they enjoy a bias reduction property similar to that of the indirect inference estimator. Monte Carlo results for a number of applications including dynamic and nonlinear panel data models, a structural auction model and two DSGE models show that the proposed estimators indeed have attractive finite sample properties.

Structured output SVM for remote sensing image classification

In the recent years, kernel methods have revealed very powerful tools in many application domains in general and in remote sensing image classification in particular. The special characteristics of remote sensing images (high dimension, few labeled samples and different noise sources) are efficiently dealt with kernel machines. In this paper, we propose the use of structured output learning to improve remote sensing image classification based on kernels. Structured output learning is concerned with the design of machine learning algorithms that not only implement input-output mapping, but also take into account the relations between output labels, thus generalizing unstructured kernel methods. We analyze the framework and introduce it to the remote sensing community. Output similarity is here encoded into SVM classifiers by modifying the model loss function and the kernel function either independently or jointly. Experiments on a very high resolution (VHR) image classification problem shows promising results and opens a wide field of research with structured output kernel methods.

The determinants of YIc's R&D activity

This paper examines the determinants of young innovative companies’ (YICs) R&D activities taking into account the autoregressive nature of innovation. Using a large longitudinal dataset comprising Spanish manufacturing firms over the period 1990-2008, we find that previous R&D experience is a fundamental determinant for mature and young firms, albeit to a smaller extent in the case of the YICs, suggesting that their innovation behaviour is less persistent and more erratic. Moreover, our results suggest that firm and market characteristics play a distinct role in boosting the innovation activity of firms of different age. In particular, while market concentration and the degree of product diversification are found to be important in boosting R&D activities in the sub-sample of mature firms only, YICs’ spending on R&D appears to be more sensitive to demand-pull variables, suggesting the presence of credit constraints. These results have been obtained using a recently proposed dynamic type-2 tobit estimator, which accounts for individual effects and efficiently handles the initial conditions problem.

The MIGCLIM R package - seamless integration of dispersal constraints into projections of species distribution models

The MIGCLIM R package is a function library for the open source R software that enables the implementation of species-specific dispersal constraints into projections of species distribution models under environmental change and/or landscape fragmentation scenarios. The model is based on a cellular automaton and the basic modeling unit is a cell that is inhabited or not. Model parameters include dispersal distance and kernel, long distance dispersal, barriers to dispersal, propagule production potential and habitat invasibility. The MIGCLIM R package has been designed to be highly flexible in the parameter values it accepts, and to offer good compatibility with existing species distribution modeling software. Possible applications include the projection of future species distributions under environmental change conditions and modeling the spread of invasive species.

Hierarchical conditional random fields for parts based models matching

A parts based model is a parametrization of an object class using a collection of landmarks following the object structure. The matching of parts based models is one of the problems where pairwise Conditional Random Fields have been successfully applied. The main reason of their effectiveness is tractable inference and learning due to the simplicity of involved graphs, usually trees. However, these models do not consider possible patterns of statistics among sets of landmarks, and thus they sufffer from using too myopic information. To overcome this limitation, we propoese a novel structure based on a hierarchical Conditional Random Fields, which we explain in the first part of this memory. We build a hierarchy of combinations of landmarks, where matching is performed taking into account the whole hierarchy. To preserve tractable inference we effectively sample the label set. We test our method on facial feature selection and human pose estimation on two challenging datasets: Buffy and MultiPIE. In the second part of this memory, we present a novel approach to multiple kernel combination that relies on stacked classification. This method can be used to evaluate the landmarks of the parts-based model approach. Our method is based on combining responses of a set of independent classifiers for each individual kernel. Unlike earlier approaches that linearly combine kernel responses, our approach uses them as inputs to another set of classifiers. We will show that we outperform state-of-the-art methods on most of the standard benchmark datasets.

Reduction theorems for operators on the cones of monotone functions

For a quasilinear operator on the semiaxis a reduction theorem is proved on the cones of monotone functions in Lp - Lq setting for 0 < q < ∞, 1<= p < ∞. The case 0 < p < 1 is also studied for operators with additional properties. In particular, we obtain critera for three-weight inequalities for the Hardy-type operators with Oinarov' kernel on monotone functions in the case 0 < q < p <= 1.

Synchronization of EEG: bivariate and multivariate measures.

Synchronization behavior of electroencephalographic (EEG) signals is important for decoding information processing in the human brain. Modern multichannel EEG allows a transition from traditional measurements of synchronization in pairs of EEG signals to whole-brain synchronization maps. The latter can be based on bivariate measures (BM) via averaging over pair-wise values or, alternatively, on multivariate measures (MM), which directly ascribe a single value to the synchronization in a group. In order to compare BM versus MM, we applied nine different estimators to simulated multivariate time series with known parameters and to real EEGs.We found widespread correlations between BM and MM, which were almost frequency-independent for all the measures except coherence. The analysis of the behavior of synchronization measures in simulated settings with variable coupling strength, connection probability, and parameter mismatch showed that some of them, including S-estimator, S-Renyi, omega, and coherence, aremore sensitive to linear interdependences,while others, like mutual information and phase locking value, are more responsive to nonlinear effects. Onemust consider these properties together with the fact thatMM are computationally less expensive and, therefore, more efficient for the large-scale data sets than BM while choosing a synchronization measure for EEG analysis.

Estadítiques de crides al sistema en plataforma GNU/Linux

Aquest projecte es basa en la modificació del kernel (nucli) del sistema operatiu GNU/Linux per dotar-lo de la capacitat d'extreure estadístiques de les crides al sistema (syscalls). A partir de la compilació i instal·lació d'un nou nucli es registra la informació del nombre de vegades i la freqüència amb què es fan aquestes crides al sistema, i posteriorment es representa en un informe d'estadístiques explicatives.

Actualització tecnológica : una evolució cap al Green Computing

El projecte que es presenta a continuació és una planificació de migració de servidors físics a un entorn virtualitzat, allà on sigui possible. A més s'ha plantejat una renovació tecnològica de tot el parc de servidors per estalviar diners en el manteniment i en el consum d'energia.La solució de virtualització es buscarà que sigui programari lliure.

Parametric estimation of pulse arrival time: a robust approach to pulse wave velocity.

Pulse wave velocity (PWV) is a surrogate of arterial stiffness and represents a non-invasive marker of cardiovascular risk. The non-invasive measurement of PWV requires tracking the arrival time of pressure pulses recorded in vivo, commonly referred to as pulse arrival time (PAT). In the state of the art, PAT is estimated by identifying a characteristic point of the pressure pulse waveform. This paper demonstrates that for ambulatory scenarios, where signal-to-noise ratios are below 10 dB, the performance in terms of repeatability of PAT measurements through characteristic points identification degrades drastically. Hence, we introduce a novel family of PAT estimators based on the parametric modeling of the anacrotic phase of a pressure pulse. In particular, we propose a parametric PAT estimator (TANH) that depicts high correlation with the Complior(R) characteristic point D1 (CC = 0.99), increases noise robustness and reduces by a five-fold factor the number of heartbeats required to obtain reliable PAT measurements.