Fast Learning by Bounding Likelihoods in Sigmoid Type Belief Networks

Sigmoid type belief networks, a class of probabilistic neural networks, provide a natural framework for compactly representing probabilistic information in a variety of unsupervised and supervised learning problems. Often the parameters used in these networks need to be learned from examples. Unfortunately, estimating the parameters via exact probabilistic calculations (i.e, the EM-algorithm) is intractable even for networks with fairly small numbers of hidden units. We propose to avoid the infeasibility of the E step by bounding likelihoods instead of computing them exactly. We introduce extended and complementary representations for these networks and show that the estimation of the network parameters can be made fast (reduced to quadratic optimization) by performing the estimation in either of the alternative domains. The complementary networks can be used for continuous density estimation as well.

Contrasting Seasonal Survivorship of Two Migratory Songbirds Wintering in Threatened Mangrove Forests

Long-distance migrants wintering in tropical regions face a number of critical conservation threats throughout their lives, but seasonal estimates of key demographic parameters such as winter survival are rare. Using mist-netting-based mark-recapture data collected in coastal Costa Rica over a six-year period, we examined variation in within- and between-winter survivorship of the Prothonotary Warbler (Protonotaria citrea; 753 young and 376 adults banded), a declining neotropical habitat specialist that depends on threatened mangrove forests during the nonbreeding season. We derived parallel seasonal survivorship estimates for the Northern Waterthrush (Seiurus noveboracensis; 564 young and 93 adults banded), a cohabitant mangrove specialist that has not shown the same population decline in North America, to assess whether contrasting survivorship might contribute to the observed differences in the species’ population trajectories. Although average annual survival probability was relatively similar between the two species for both young and adult birds, monthly estimates indicated that relative to Northern Waterthrush, Prothonotary Warblers exhibited: greater interannual variation in survivorship, especially within winters; greater variation in survivorship among the three study sites; lower average between-winter survivorship, particularly among females, and; a sharp decline in between-winter survivorship from 2003 to 2009 for both age groups and both sexes. Rather than identifying one seasonal vital rate as a causal factor of Prothonotary Warbler population declines, our species comparison suggests that the combination of variable within-winter survival with decreasing between-winter survival demands a multi-seasonal approach to the conservation of this and other tropical-wintering migrants.

An Earth Observation Land Data Assimilation System (EO-LDAS)

Current methods for estimating vegetation parameters are generally sub-optimal in the way they exploit information and do not generally consider uncertainties. We look forward to a future where operational dataassimilation schemes improve estimates by tracking land surface processes and exploiting multiple types of observations. Dataassimilation schemes seek to combine observations and models in a statistically optimal way taking into account uncertainty in both, but have not yet been much exploited in this area. The EO-LDAS scheme and prototype, developed under ESA funding, is designed to exploit the anticipated wealth of data that will be available under GMES missions, such as the Sentinel family of satellites, to provide improved mapping of land surface biophysical parameters. This paper describes the EO-LDAS implementation, and explores some of its core functionality. EO-LDAS is a weak constraint variational dataassimilationsystem. The prototype provides a mechanism for constraint based on a prior estimate of the state vector, a linear dynamic model, and EarthObservationdata (top-of-canopy reflectance here). The observation operator is a non-linear optical radiative transfer model for a vegetation canopy with a soil lower boundary, operating over the range 400 to 2500 nm. Adjoint codes for all model and operator components are provided in the prototype by automatic differentiation of the computer codes. In this paper, EO-LDAS is applied to the problem of daily estimation of six of the parameters controlling the radiative transfer operator over the course of a year (> 2000 state vector elements). Zero and first order process model constraints are implemented and explored as the dynamic model. The assimilation estimates all state vector elements simultaneously. This is performed in the context of a typical Sentinel-2 MSI operating scenario, using synthetic MSI observations simulated with the observation operator, with uncertainties typical of those achieved by optical sensors supposed for the data. The experiments consider a baseline state vector estimation case where dynamic constraints are applied, and assess the impact of dynamic constraints on the a posteriori uncertainties. The results demonstrate that reductions in uncertainty by a factor of up to two might be obtained by applying the sorts of dynamic constraints used here. The hyperparameter (dynamic model uncertainty) required to control the assimilation are estimated by a cross-validation exercise. The result of the assimilation is seen to be robust to missing observations with quite large data gaps.

Bayesian parameter estimation for latent Markov random fields and social networks

Undirected graphical models are widely used in statistics, physics and machine vision. However Bayesian parameter estimation for undirected models is extremely challenging, since evaluation of the posterior typically involves the calculation of an intractable normalising constant. This problem has received much attention, but very little of this has focussed on the important practical case where the data consists of noisy or incomplete observations of the underlying hidden structure. This paper specifically addresses this problem, comparing two alternative methodologies. In the first of these approaches particle Markov chain Monte Carlo (Andrieu et al., 2010) is used to efficiently explore the parameter space, combined with the exchange algorithm (Murray et al., 2006) for avoiding the calculation of the intractable normalising constant (a proof showing that this combination targets the correct distribution in found in a supplementary appendix online). This approach is compared with approximate Bayesian computation (Pritchard et al., 1999). Applications to estimating the parameters of Ising models and exponential random graphs from noisy data are presented. Each algorithm used in the paper targets an approximation to the true posterior due to the use of MCMC to simulate from the latent graphical model, in lieu of being able to do this exactly in general. The supplementary appendix also describes the nature of the resulting approximation.

Late Pleistocene climate change and the global expansion of anatomically modern humans

The extent to which past climate change has dictated the pattern and timing of the out-of-Africa expansion by anatomically modern humans is currently unclear [Stewart JR, Stringer CB (2012) Science 335:1317–1321]. In particular, the incompleteness of the fossil record makes it difficult to quantify the effect of climate. Here, we take a different approach to this problem; rather than relying on the appearance of fossils or archaeological evidence to determine arrival times in different parts of the world, we use patterns of genetic variation in modern human populations to determine the plausibility of past demographic parameters. We develop a spatially explicit model of the expansion of anatomically modern humans and use climate reconstructions over the past 120 ky based on the Hadley Centre global climate model HadCM3 to quantify the possible effects of climate on human demography. The combinations of demographic parameters compatible with the current genetic makeup of worldwide populations indicate a clear effect of climate on past population densities. Our estimates of this effect, based on population genetics, capture the observed relationship between current climate and population density in modern hunter–gatherers worldwide, providing supporting evidence for the realism of our approach. Furthermore, although we did not use any archaeological and anthropological data to inform the model, the arrival times in different continents predicted by our model are also broadly consistent with the fossil and archaeological records. Our framework provides the most accurate spatiotemporal reconstruction of human demographic history available at present and will allow for a greater integration of genetic and archaeological evidence.

Single-carrier frequency domain equalisation for hammerstein communication systems using complex-valued neural networks

Single-carrier (SC) block transmission with frequency-domain equalisation (FDE) offers a viable transmission technology for combating the adverse effects of long dispersive channels encountered in high-rate broadband wireless communication systems. However, for high bandwidthefficiency and high power-efficiency systems, the channel can generally be modelled by the Hammerstein system that includes the nonlinear distortion effects of the high power amplifier (HPA) at transmitter. For such nonlinear Hammerstein channels, the standard SC-FDE scheme no longer works. This paper advocates a complex-valued (CV) B-spline neural network based nonlinear SC-FDE scheme for Hammerstein channels. Specifically, We model the nonlinear HPA, which represents the CV static nonlinearity of the Hammerstein channel, by a CV B-spline neural network, and we develop two efficient alternating least squares schemes for estimating the parameters of the Hammerstein channel, including both the channel impulse response coefficients and the parameters of the CV B-spline model. We also use another CV B-spline neural network to model the inversion of the nonlinear HPA, and the parameters of this inverting B-spline model can easily be estimated using the standard least squares algorithm based on the pseudo training data obtained as a natural byproduct of the Hammerstein channel identification. Equalisation of the SC Hammerstein channel can then be accomplished by the usual one-tap linear equalisation in frequency domain as well as the inverse B-spline neural network model obtained in time domain. Extensive simulation results are included to demonstrate the effectiveness of our nonlinear SC-FDE scheme for Hammerstein channels.

Nonlinear equalization of Hammerstein OFDM systems

A practical orthogonal frequency-division multiplexing (OFDM) system can generally be modelled by the Hammerstein system that includes the nonlinear distortion effects of the high power amplifier (HPA) at transmitter. In this contribution, we advocate a novel nonlinear equalization scheme for OFDM Hammerstein systems. We model the nonlinear HPA, which represents the static nonlinearity of the OFDM Hammerstein channel, by a B-spline neural network, and we develop a highly effective alternating least squares algorithm for estimating the parameters of the OFDM Hammerstein channel, including channel impulse response coefficients and the parameters of the B-spline model. Moreover, we also use another B-spline neural network to model the inversion of the HPA’s nonlinearity, and the parameters of this inverting B-spline model can easily be estimated using the standard least squares algorithm based on the pseudo training data obtained as a byproduct of the Hammerstein channel identification. Equalization of the OFDM Hammerstein channel can then be accomplished by the usual one-tap linear equalization as well as the inverse B-spline neural network model obtained. The effectiveness of our nonlinear equalization scheme for OFDM Hammerstein channels is demonstrated by simulation results.

B-spline neural network based single-carrier frequency domain equalisation for Hammerstein channels

A practical single-carrier (SC) block transmission with frequency domain equalisation (FDE) system can generally be modelled by the Hammerstein system that includes the nonlinear distortion effects of the high power amplifier (HPA) at transmitter. For such Hammerstein channels, the standard SC-FDE scheme no longer works. We propose a novel Bspline neural network based nonlinear SC-FDE scheme for Hammerstein channels. In particular, we model the nonlinear HPA, which represents the complex-valued static nonlinearity of the Hammerstein channel, by two real-valued B-spline neural networks, one for modelling the nonlinear amplitude response of the HPA and the other for the nonlinear phase response of the HPA. We then develop an efficient alternating least squares algorithm for estimating the parameters of the Hammerstein channel, including the channel impulse response coefficients and the parameters of the two B-spline models. Moreover, we also use another real-valued B-spline neural network to model the inversion of the HPA’s nonlinear amplitude response, and the parameters of this inverting B-spline model can be estimated using the standard least squares algorithm based on the pseudo training data obtained as a byproduct of the Hammerstein channel identification. Equalisation of the SC Hammerstein channel can then be accomplished by the usual one-tap linear equalisation in frequency domain as well as the inverse Bspline neural network model obtained in time domain. The effectiveness of our nonlinear SC-FDE scheme for Hammerstein channels is demonstrated in a simulation study.

Interactions between climate change and land use change on biodiversity: attribution problems, risks, and opportunities

Global change drivers are known to interact in their effects on biodiversity, but much research to date ignores this complexity. As a consequence, there are problems in the attribution of biodiversity change to different drivers and, therefore, our ability to manage habitats and landscapes appropriately. Few studies explicitly acknowledge and account for interactive (i.e., nonadditive) effects of land use and climate change on biodiversity. One reason is that the mechanisms by which drivers interact are poorly understood. We evaluate such mechanisms, including interactions between demographic parameters, evolutionary trade-offs and synergies and threshold effects of population size and patch occupancy on population persistence. Other reasons for the lack of appropriate research are limited data availability and analytical issues in addressing interaction effects. We highlight the influence that attribution errors can have on biodiversity projections and discuss experimental designs and analytical tools suited to this challenge. Finally, we summarize the risks and opportunities provided by the existence of interaction effects. Risks include ineffective conservation management; but opportunities also arise, whereby the negative impacts of climate change on biodiversity can be reduced through appropriate land management as an adaptation measure. We hope that increasing the understanding of key mechanisms underlying interaction effects and discussing appropriate experimental and analytical designs for attribution will help researchers, policy makers, and conservation practitioners to better minimize risks and exploit opportunities provided by land use-climate change interactions.

Consumption-wealth ratio and expected stock returns: evidence from panel data

This paper investigates the role of consumption-wealth ratio on predicting future stock returns through a panel approach. We follow the theoretical framework proposed by Lettau and Ludvigson (2001), in which a model derived from a nonlinear consumer’s budget constraint is used to settle the link between consumption-wealth ratio and stock returns. Using G7’s quarterly aggregate and financial data ranging from the first quarter of 1981 to the first quarter of 2014, we set an unbalanced panel that we use for both estimating the parameters of the cointegrating residual from the shared trend among consumption, asset wealth and labor income, cay, and performing in and out-of-sample forecasting regressions. Due to the panel structure, we propose different methodologies of estimating cay and making forecasts from the one applied by Lettau and Ludvigson (2001). The results indicate that cay is in fact a strong and robust predictor of future stock return at intermediate and long horizons, but presents a poor performance on predicting one or two-quarter-ahead stock returns.

Heritabilities and genetic correlations for reproductive traits in an F2 reciprocal cross chicken population

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Parâmetros genéticos para produção de leite no dia do controle e para produção de leite até 305 dias nas primeiras lactações de vacas da raça Gir

Foram estimados parâmetros genéticos para as produções de leite no dia do controle (PLDC) e para a produção acumulada até 305 dias (P305) de 2.440 primeiras lactações de vacas da raça Gir leiteiro com partos entre 1990 e 2005. A produção acumulada até 305 dias e as produções de leite no dia do controle foram analisadas por meio de um modelo animal em análises uni e bicaracterísticas. Para as produções no dia do controle, o modelo incluiu o efeito aleatório genético direto e os efeitos fixos de grupo de contemporâneos, classe de idade da vaca ao parto e dias em lactação como co-variável (regressão linear). Para a P305, utilizou-se o mesmo modelo, excluindo-se os dias em lactação e utilizando-se a idade da vaca ao primeiro parto como co-variável (regressão linear e quadrática). O grupo de contemporâneos (GC) foi definido como rebanho, ano e estação do controle para produções no dia do controle e rebanho, ano e época de parto para P305. Aplicou-se a restrição de que cada grupo de contemporâneo deveria conter no mínimo três observações no caso das produções no dia do controle e cinco observações para P305. Os componentes de variância foram estimados pelo método de máxima verossimilhança restrita. Os valores maiores de variâncias fenotípicas foram observados no meio da lactação e os das variâncias aditivas no início da lactação. As estimativas de herdabilidade obtidas pelas análises unicaracterísticas oscilaram entre 0,13 e 0,36, enquanto as obtidas por análises bi-características oscilaram entre 0,14 e 0,34. Para as duas análises, essas estimativas foram maiores nos primeiros e menores nos últimos controles. As correlações genéticas entre as produções no dia do controle foram elevadas e positivas e maiores entre os controles mais próximos. Os resultados sugerem que as produções no dia do controle podem ser utilizadas como critério de seleção para produção de leite da raça Gir.

Análise genética de escores de avaliação visual de bovinos com modelos bayesianos de limiar e linear

O objetivo deste trabalho foi comparar as estimativas de parâmetros genéticos obtidas em análises bayesianas uni-característica e bi-característica, em modelo animal linear e de limiar, considerando-se as características categóricas morfológicas de bovinos da raça Nelore. Os dados de musculosidade, estrutura física e conformação foram obtidos entre 2000 e 2005, em 3.864 animais de 13 fazendas participantes do Programa Nelore Brasil. Foram realizadas análises bayesianas uni e bi-características, em modelos de limiar e linear. de modo geral, os modelos de limiar e linear foram eficientes na estimação dos parâmetros genéticos para escores visuais em análises bayesianas uni-características. Nas análises bi-características, observou-se que: com utilização de dados contínuos e categóricos, o modelo de limiar proporcionou estimativas de correlação genética de maior magnitude do que aquelas do modelo linear; e com o uso de dados categóricos, as estimativas de herdabilidade foram semelhantes. A vantagem do modelo linear foi o menor tempo gasto no processamento das análises. Na avaliação genética de animais para escores visuais, o uso do modelo de limiar ou linear não influenciou a classificação dos animais, quanto aos valores genéticos preditos, o que indica que ambos os modelos podem ser utilizados em programas de melhoramento genético.

Fuzzy subset approach in coupled population dynamics of blowflies

This paper is a study on the population dynamics of blowflies employing a density-dependent. non-linear mathematical model and a coupled population formalism. In this Study, we investigated the coupled population dynamics applying fuzzy subsets to model the Population trajectory. analyzing demographic parameters such as fecundity, Survival, and migration. The main results suggest different possibilities in terms of dynamic behavior produced by migration in coupled Populations between distinct environments and the rescue effect generated by the connection between populations. It was possible to conclude that environmental heterogeneity can play an important role in blowfly metapopulation systems. The implications of these results for population dynamics of blowflies are discussed.

Sex ratio and dynamic behavior in populations of the exotic blowfly Chrysomya albiceps (Diptera, Calliphoridae)

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)