Generalized KPCA by adaptive rules in feature space

Principal component analysis (PCA) is well recognized in dimensionality reduction, and kernel PCA (KPCA) has also been proposed in statistical data analysis. However, KPCA fails to detect the nonlinear structure of data well when outliers exist. To reduce this problem, this paper presents a novel algorithm, named iterative robust KPCA (IRKPCA). IRKPCA works well in dealing with outliers, and can be carried out in an iterative manner, which makes it suitable to process incremental input data. As in the traditional robust PCA (RPCA), a binary field is employed for characterizing the outlier process, and the optimization problem is formulated as maximizing marginal distribution of a Gibbs distribution. In this paper, this optimization problem is solved by stochastic gradient descent techniques. In IRKPCA, the outlier process is in a high-dimensional feature space, and therefore kernel trick is used. IRKPCA can be regarded as a kernelized version of RPCA and a robust form of kernel Hebbian algorithm. Experimental results on synthetic data demonstrate the effectiveness of IRKPCA. © 2010 Taylor & Francis.

Data compilation on the biological response to ocean acidification: environmental and experimental context of data sets and related literature

The exponential growth of studies on the biological response to ocean acidification over the last few decades has generated a large amount of data. To facilitate data comparison, a data compilation hosted at the data publisher PANGAEA was initiated in 2008 and is updated on a regular basis (doi:10.1594/PANGAEA.149999). By January 2015, a total of 581 data sets (over 4 000 000 data points) from 539 papers had been archived. Here we present the developments of this data compilation five years since its first description by Nisumaa et al. (2010). Most of study sites from which data archived are still in the Northern Hemisphere and the number of archived data from studies from the Southern Hemisphere and polar oceans are still relatively low. Data from 60 studies that investigated the response of a mix of organisms or natural communities were all added after 2010, indicating a welcomed shift from the study of individual organisms to communities and ecosystems. The initial imbalance of considerably more data archived on calcification and primary production than on other processes has improved. There is also a clear tendency towards more data archived from multifactorial studies after 2010. For easier and more effective access to ocean acidification data, the ocean acidification community is strongly encouraged to contribute to the data archiving effort, and help develop standard vocabularies describing the variables and define best practices for archiving ocean acidification data.

A figuratividade nas histórias em quadrinhos: uma análise das construções metafóricas e metonímicas em V de vingança

The object of this study is the construction of metaphor and metonymy in comics. This work is inserted in the field of Embodied Cognitive Linguistics, specifically based on the Neural Theory of Language (FELDMAN, 2006) and, consistent with this theoretical and methodological framework, the notions of categorization (LAKOFF & JOHNSON, 1999), embodiment (GIBBS, 2005), figurativity (GIBBS, 1994; BERGEN, 2005), and mental simulation (BARSALOU, 1999; FELDMAN, 2006) have also been used. The hypothesis defended is that the construction of figurativity in texts consisting of verbal and nonverbal mechanisms is linked to the activation of neural structures related to our actions and perceptions. Thus, language is considered a cognitive faculty connected to the brain apparatus and to bodily experiences, in such a way that it provides samples of the continuous process of meaning (re)construction performed by the reader, whom (re)defines his or her views about the world as certain neural networks are (or stop being) activated during linguistic processing. The data obtained during the analysys shows that, as regards comics, the act of reading together the graphics and verbal language seems to have an important role in the construction of figurativity, including cases of metaphors which are metonymically motivated. These preliminary conclusions were drawn from the data analysis taken from V de Vingança (MOORE; LLOYD, 2006). The corpus study was guided by the methodology of introspection, i.e., the individual analysis of linguistic aspects as manifested in one's own cognition (TALMY, 2005).

Caracterização físico-química de sistemas micelares

The process of adsorption and micellization of the surfactants sodium dodecyl sulfate, dodecylammonium chloride and hexaethylene glycol mono-n-dodecyl ether in water-air interface has been studied using measurements of surface tension in aqueous media and NaCl 0.1 mol/L in temperatures of 25, 33 and 40 °C. From these data, critical micelle concentrations and thermodynamic parameters of micellization and adsorption were determined in order to elucidate the behaviors of micellization and adsorption for these surfactants in the proposed medium. For the determination of the thermodynamic parameters of adsorption we utilized the equations of isotherms of Langmuir and Gibbs. Γmáx values determined by the different equations were correlated to the explanation of results. Temperature and salinity were analyzed in terms of their influence on the micellization and adsorption process, and the results were explained based on intermolecular interactions. The values of Gmic have confirmed that the micelle formation for the surfactants studied occurs spontaneously

Estudo teórico de intermediários tetraédricos acidez / basicidade e estereosseletividade enzimáticos

The present work aimed first, the theoretical study of tetrahedral intermediate stability formed from carbonyl addition reactions using the second (MP2) and third (MP3) order Møller–Plesset perturbation theory. Linear correlations between electronic energy difference of reactions with Wiberg Indexes and C-O bond lengths were obtained, and was observed that the stability of adducts formed depends directly of electronic density involved between these atoms. The knowing of electronic parameters of these structures has an important hole due to the large use on reactions that in his course forms this tetrahedral intermediate. Employing the ONIOM (B3LYP:AMBER) methodology, was evaluated the stereoselectivity of a enzymatic reaction between CAL B enzyme and a long chain ester. In this study, were obtained the electronic energies of ground state and intermediate state of transesterification rate-determing step from two possible proquirals faces Re and Si. The objective was study the enantioselectivity of CAL B and rationalizes it using quantum theory of atoms in molecules (QTAIM). A theoretical study employing inorganic compounds was performed using ab initio CBS-QB3 method aiming to find a link between thermodynamic and equilibrium involving acids and bases. The results observed showed an excellent relationship between difference in Gibbs free energy, ΔG of acid dissociation reaction and ΔG of hydrolysis reaction of the corresponding conjugate base. It was also observed, a relationship between ΔG of hydrolysis reaction of conjugate acids and their corresponding atomic radius showing that stability plays an important role in hydrolysis reactions. The importance of solvation in acid/base behavior when compared to theoretical and experimental ΔG´s also was evaluated.

Síntese e caracterização de materiais mesoporosos para a captura de CO2: influência do óxido de níquel

Several materials are currently under study for the CO2 capture process, like the metal oxides and mixed metal oxides, zeolites, carbonaceous materials, metal-organic frameworks (MOF's) organosilica and modified silica surfaces. In this work, evaluated the adsorption capacity of CO2 in mesoporous materials of different structures, such as MCM-48 and SBA- 15 without impregnating and impregnated with nickel in the proportions 5 %, 10 % and 20 % (m/m), known as 5Ni-MCM-48, 10Ni-MCM-48, 20Ni-MCM-48 and 5Ni-SBA-15, 10NiSBA-15, 20Ni-SBA-15. The materials were characterized by means of X-ray diffraction (XRD), thermal analysis (TG and DTG), Fourier transform infrared spectroscopy (FT-IR), N2 adsorption and desorption (BET) and scanning electron microscopy (SEM) with EDS. The adsorption process was performed varying the pressure of 100 - 4000 kPa and keeping the temperature constant and equal to 298 K. At a pressure of 100 kPa, higher concentrations of adsorption occurred for the materials 5Ni-MCM-48 (0.795 mmol g-1 ) and SBA-15 (0.914 mmol g-1 ) is not impregnated, and at a pressure of 4000 kPa for MCM-48 materials (14.89 mmol g-1) and SBA-15 (9.97 mmol g-1) not impregnated. The results showed that the adsorption capacity varies positively with the specific area, however, has a direct dependency on the type and geometry of the porous structure of channels. The data were fitted using the Langmuir and Freundlich models and were evaluated thermodynamic parameters Gibbs free energy and entropy of the adsorption system

Geochemical and isotopic composition of Bransfield Basin Lavas collected during SONNE cruise SO155

Bransfield Basin is an actively extending marginal basin separating the inactive South Shetland arc from the northern Antarctic Peninsula. Rift-related volcanism is widespread throughout the central Bransfield Basin, but the wider eastern Bransfield Basin was previously unsampled. Lavas recovered from the eastern subbasin form three distinct groups: (1) Bransfield Group has moderate large-ion lithophile element (LILE) enrichment relative to normal mid-ocean ridge basalt (NMORB), (2) Gibbs Group has strong LILE enrichment and is restricted to a relic seamount interpreted as part of the South Shetland arc, and (3) fresh alkali basalt was recovered from the NE part of the basin near Spanish Rise. The subduction-related component in Bransfield and Gibbs Group lavas is a LILE-rich fluid with radiogenic Sr, Nd, and Pb isotope compositions derived predominantly from subducting sediment. These lavas can be modeled as melts from Pacific MORB source mantle contaminated by up to 5% of the subduction-related component. They further reveal that Pacific mantle, rather than South Atlantic mantle, has underlain Bransfield Basin since 3 Ma. Magma productivity decreases abruptly east of Bridgeman Rise, and lavas with the least subduction component outcrop at that end. Both the eastward decrease in subduction component and occurrence of young alkali basalts require that subduction-modified mantle generated during the lifetime of the South Shetland arc has been progressively removed from NE to SW. This is inconsistent with previous models suggesting continued slow subduction at the South Shetland Trench but instead favors models in which the South Scotia Ridge fault has propagated westward since 3 Ma generating transtension across the basin.

Mg/Ca and Sr/Ca ratios of planktonic foraminifera from North Atlantic surface sediments

Reliable temperature estimates from both surface and subsurface ocean waters are needed to reconstruct past upper water column temperature gradients and past oceanic heat content. This work examines the relationships between trace element ratios in fossil shells and seawater temperature for surface-dwelling foraminifera species, Globigerinoides ruber (white) and Globigerina bulloides, and deep-dwelling species, Globorotalia inflata, Globorotalia truncatulinoides (dextral and sinistral) and Pulleniatina obliquiloculata. Mg/Ca and Sr/Ca ratios in shells picked in 29 modern core tops from the North Atlantic Ocean are calibrated using calculated isotopic temperatures. Mg/Ca ratios on G. ruber and G. bulloides agree with published data and relationships. For deep-dwelling species, Mg/Ca calibration follows the equation Mg/Ca = 0.78 (±0.04) * exp (0.051 (±0.003) * T) with a significant correlation coefficient of R**2 = 0.74. Moreover, there is no significant difference between the different deep-dwellers analyzed. For the Sr/Ca ratio, the surface dwellers and P. obliquiloculata do not record any temperature dependence. For the Globorotalia species, the thermo dependence of Sr/Ca ratio can be described by a single linear relationship: Sr/Ca = (0.0182 (±0.001) * T) + 1.097 (±0.018), R**2 = 0.85. Temperature estimates with a 1 sigma error of ±2.0°C and ±1.3°C can be derived from the Mg/Ca and Sr/Ca ratios, respectively, as long as the Sr geochemistry in the ocean has been constant through time.

Stable isotope ratios of planktonic foraminifera from North Atlantic surface sediments

We measured the oxygen isotopic composition of the deep-dwelling foraminiferal species G. inflata, G. truncatulinoides dextral and sinistral, and P. obliquiloculata in 29 modern core tops raised from the North Atlantic Ocean. We compared calculated isotopic temperatures with atlas temperatures and defined ecological models for each species. G. inflata and G. truncatulinoides live preferentially at the base of the seasonal thermocline. Under temperature stress, i.e., when the base of the seasonal thermocline is warmer than 16°C, G. inflata and G. truncatulinoides live deeper in the main thermocline. P. obliquiloculata inhabits the seasonal thermocline in warm regions. We tested our model using 10 cores along the Mauritanian upwelling and show that the comparison of d18O variations registered by the surficial species G. ruber and G. bulloides and the deep-dwelling species G. inflata evidences significant glacial-interglacial shifts of the Mauritanian upwelling cells.

Seawater carbonate chemistry and processes during experiments with Emiliania huxleyi, 2008

Ocean acidification in response to rising atmospheric CO2 partial pressures is widely expected to reduce calcification by marine organisms. From the mid-Mesozoic, coccolithophores have been major calcium carbonate producers in the world's oceans, today accounting for about a third of the total marine CaCO3 production. Here, we present laboratory evidence that calcification and net primary production in the coccolithophore species Emiliania huxleyi are significantly increased by high CO2 partial pressures. Field evidence from the deep ocean is consistent with these laboratory conclusions, indicating that over the past 220 years there has been a 40% increase in average coccolith mass. Our findings show that coccolithophores are already responding and will probably continue to respond to rising atmospheric CO2 partial pressures, which has important implications for biogeochemical modeling of future oceans and climate.

Managing the multiplicities of graduate level education! Are Master’s students’ expectations matched by their experiences?

With rapid increases in student fees reflecting moves towards a QUASI Market model of Higher Education in the UK and across much of the Western World[1], many universities find themselves having to meet progressively higher levels of student expectations[2]. This is particularly the case at undergraduate level, where increases in fees over the past decade have far exceeded inflation. Yet with so much attention on ‘consumer savvy’ undergraduates, the question of whether Master’s level students’ expectations are matched by their experiences is one which remains largely unanswered. Grounded in an empirically grounded approach to learning and teaching developed by the paper authors[3], this paper sets out to being to answer this question. In doing so it makes a distinctive contribution to debates about graduate level engineering education and concludes with a number of recommendations. Discussion: The ‘MSc: Managing Expectations’ Project analyses the expectations and experiences of Graduate level Engineering Management Students over a two year period. Focusingon the ‘student experience’, three main concepts are identified as being particular relevant to enhancing learning [3]: Relationships: Variety: Synergy. Relationships: Based on empirical research, the significance of Relationships within the academic environment is discussed with particular attention being paid to the value of students’ social and academic support networks, including academic tutoring. Variety: Grounded in a statistical analysis of ‘engagement data’ together with survey and interview findings, the concept of variety critically examines students’ perspectives and experiencesof different approaches to learning and teaching. Synergy: Possibly the most important concept discussed within this paper, the need for constructively aligned curriculum is extended to reflect the students’ apriori knowledge and experienceas well as employer and societal demands and expectations. The conclusion brings the different concepts within the discussion together, providing a set of practical recommendations for colleagues working both at graduate and undergraduate level. References 1.Gibbs, P. (2001) "Higher education as a market: a problem or solution?." Studies in Higher Education 26. 1. pp. 85-94. 2.Tricker, T., (2005) Student Expectations-How do we measure up. University of Sheffield. Available from: http://www.persons.org.uk/tricker%20paper.pdf Accessed 9/10/14 3.Clark, R. & Andrews, J. (2014). Relationships, Variety & Synergy [RVS]: The Vital Ingredients for Scholarship in Engineering Education? A Case-Study. European Journal of Engineering Education. 39.6. pp. 585-600.

Bayesian Inference in Large-scale Problems

Many modern applications fall into the category of "large-scale" statistical problems, in which both the number of observations n and the number of features or parameters p may be large. Many existing methods focus on point estimation, despite the continued relevance of uncertainty quantification in the sciences, where the number of parameters to estimate often exceeds the sample size, despite huge increases in the value of n typically seen in many fields. Thus, the tendency in some areas of industry to dispense with traditional statistical analysis on the basis that "n=all" is of little relevance outside of certain narrow applications. The main result of the Big Data revolution in most fields has instead been to make computation much harder without reducing the importance of uncertainty quantification. Bayesian methods excel at uncertainty quantification, but often scale poorly relative to alternatives. This conflict between the statistical advantages of Bayesian procedures and their substantial computational disadvantages is perhaps the greatest challenge facing modern Bayesian statistics, and is the primary motivation for the work presented here.

Two general strategies for scaling Bayesian inference are considered. The first is the development of methods that lend themselves to faster computation, and the second is design and characterization of computational algorithms that scale better in n or p. In the first instance, the focus is on joint inference outside of the standard problem of multivariate continuous data that has been a major focus of previous theoretical work in this area. In the second area, we pursue strategies for improving the speed of Markov chain Monte Carlo algorithms, and characterizing their performance in large-scale settings. Throughout, the focus is on rigorous theoretical evaluation combined with empirical demonstrations of performance and concordance with the theory.

One topic we consider is modeling the joint distribution of multivariate categorical data, often summarized in a contingency table. Contingency table analysis routinely relies on log-linear models, with latent structure analysis providing a common alternative. Latent structure models lead to a reduced rank tensor factorization of the probability mass function for multivariate categorical data, while log-linear models achieve dimensionality reduction through sparsity. Little is known about the relationship between these notions of dimensionality reduction in the two paradigms. In Chapter 2, we derive several results relating the support of a log-linear model to nonnegative ranks of the associated probability tensor. Motivated by these findings, we propose a new collapsed Tucker class of tensor decompositions, which bridge existing PARAFAC and Tucker decompositions, providing a more flexible framework for parsimoniously characterizing multivariate categorical data. Taking a Bayesian approach to inference, we illustrate empirical advantages of the new decompositions.

Latent class models for the joint distribution of multivariate categorical, such as the PARAFAC decomposition, data play an important role in the analysis of population structure. In this context, the number of latent classes is interpreted as the number of genetically distinct subpopulations of an organism, an important factor in the analysis of evolutionary processes and conservation status. Existing methods focus on point estimates of the number of subpopulations, and lack robust uncertainty quantification. Moreover, whether the number of latent classes in these models is even an identified parameter is an open question. In Chapter 3, we show that when the model is properly specified, the correct number of subpopulations can be recovered almost surely. We then propose an alternative method for estimating the number of latent subpopulations that provides good quantification of uncertainty, and provide a simple procedure for verifying that the proposed method is consistent for the number of subpopulations. The performance of the model in estimating the number of subpopulations and other common population structure inference problems is assessed in simulations and a real data application.

In contingency table analysis, sparse data is frequently encountered for even modest numbers of variables, resulting in non-existence of maximum likelihood estimates. A common solution is to obtain regularized estimates of the parameters of a log-linear model. Bayesian methods provide a coherent approach to regularization, but are often computationally intensive. Conjugate priors ease computational demands, but the conjugate Diaconis--Ylvisaker priors for the parameters of log-linear models do not give rise to closed form credible regions, complicating posterior inference. In Chapter 4 we derive the optimal Gaussian approximation to the posterior for log-linear models with Diaconis--Ylvisaker priors, and provide convergence rate and finite-sample bounds for the Kullback-Leibler divergence between the exact posterior and the optimal Gaussian approximation. We demonstrate empirically in simulations and a real data application that the approximation is highly accurate, even in relatively small samples. The proposed approximation provides a computationally scalable and principled approach to regularized estimation and approximate Bayesian inference for log-linear models.

Another challenging and somewhat non-standard joint modeling problem is inference on tail dependence in stochastic processes. In applications where extreme dependence is of interest, data are almost always time-indexed. Existing methods for inference and modeling in this setting often cluster extreme events or choose window sizes with the goal of preserving temporal information. In Chapter 5, we propose an alternative paradigm for inference on tail dependence in stochastic processes with arbitrary temporal dependence structure in the extremes, based on the idea that the information on strength of tail dependence and the temporal structure in this dependence are both encoded in waiting times between exceedances of high thresholds. We construct a class of time-indexed stochastic processes with tail dependence obtained by endowing the support points in de Haan's spectral representation of max-stable processes with velocities and lifetimes. We extend Smith's model to these max-stable velocity processes and obtain the distribution of waiting times between extreme events at multiple locations. Motivated by this result, a new definition of tail dependence is proposed that is a function of the distribution of waiting times between threshold exceedances, and an inferential framework is constructed for estimating the strength of extremal dependence and quantifying uncertainty in this paradigm. The method is applied to climatological, financial, and electrophysiology data.

The remainder of this thesis focuses on posterior computation by Markov chain Monte Carlo. The Markov Chain Monte Carlo method is the dominant paradigm for posterior computation in Bayesian analysis. It has long been common to control computation time by making approximations to the Markov transition kernel. Comparatively little attention has been paid to convergence and estimation error in these approximating Markov Chains. In Chapter 6, we propose a framework for assessing when to use approximations in MCMC algorithms, and how much error in the transition kernel should be tolerated to obtain optimal estimation performance with respect to a specified loss function and computational budget. The results require only ergodicity of the exact kernel and control of the kernel approximation accuracy. The theoretical framework is applied to approximations based on random subsets of data, low-rank approximations of Gaussian processes, and a novel approximating Markov chain for discrete mixture models.

Data augmentation Gibbs samplers are arguably the most popular class of algorithm for approximately sampling from the posterior distribution for the parameters of generalized linear models. The truncated Normal and Polya-Gamma data augmentation samplers are standard examples for probit and logit links, respectively. Motivated by an important problem in quantitative advertising, in Chapter 7 we consider the application of these algorithms to modeling rare events. We show that when the sample size is large but the observed number of successes is small, these data augmentation samplers mix very slowly, with a spectral gap that converges to zero at a rate at least proportional to the reciprocal of the square root of the sample size up to a log factor. In simulation studies, moderate sample sizes result in high autocorrelations and small effective sample sizes. Similar empirical results are observed for related data augmentation samplers for multinomial logit and probit models. When applied to a real quantitative advertising dataset, the data augmentation samplers mix very poorly. Conversely, Hamiltonian Monte Carlo and a type of independence chain Metropolis algorithm show good mixing on the same dataset.

Advances in Dynamic Modeling and Computation for Count Data

A class of multi-process models is developed for collections of time indexed count data. Autocorrelation in counts is achieved with dynamic models for the natural parameter of the binomial distribution. In addition to modeling binomial time series, the framework includes dynamic models for multinomial and Poisson time series. Markov chain Monte Carlo (MCMC) and Po ́lya-Gamma data augmentation (Polson et al., 2013) are critical for fitting multi-process models of counts. To facilitate computation when the counts are high, a Gaussian approximation to the P ́olya- Gamma random variable is developed.

Three applied analyses are presented to explore the utility and versatility of the framework. The first analysis develops a model for complex dynamic behavior of themes in collections of text documents. Documents are modeled as a “bag of words”, and the multinomial distribution is used to characterize uncertainty in the vocabulary terms appearing in each document. State-space models for the natural parameters of the multinomial distribution induce autocorrelation in themes and their proportional representation in the corpus over time.

The second analysis develops a dynamic mixed membership model for Poisson counts. The model is applied to a collection of time series which record neuron level firing patterns in rhesus monkeys. The monkey is exposed to two sounds simultaneously, and Gaussian processes are used to smoothly model the time-varying rate at which the neuron’s firing pattern fluctuates between features associated with each sound in isolation.

The third analysis presents a switching dynamic generalized linear model for the time-varying home run totals of professional baseball players. The model endows each player with an age specific latent natural ability class and a performance enhancing drug (PED) use indicator. As players age, they randomly transition through a sequence of ability classes in a manner consistent with traditional aging patterns. When the performance of the player significantly deviates from the expected aging pattern, he is identified as a player whose performance is consistent with PED use.

All three models provide a mechanism for sharing information across related series locally in time. The models are fit with variations on the P ́olya-Gamma Gibbs sampler, MCMC convergence diagnostics are developed, and reproducible inference is emphasized throughout the dissertation.

Multi-View Weighted Network

Extensive investigation has been conducted on network data, especially weighted network in the form of symmetric matrices with discrete count entries. Motivated by statistical inference on multi-view weighted network structure, this paper proposes a Poisson-Gamma latent factor model, not only separating view-shared and view-specific spaces but also achieving reduced dimensionality. A multiplicative gamma process shrinkage prior is implemented to avoid over parameterization and efficient full conditional conjugate posterior for Gibbs sampling is accomplished. By the accommodating of view-shared and view-specific parameters, flexible adaptability is provided according to the extents of similarity across view-specific space. Accuracy and efficiency are tested by simulated experiment. An application on real soccer network data is also proposed to illustrate the model.