Regularisation of mixture density networks

Mixture Density Networks are a principled method to model conditional probability density functions which are non-Gaussian. This is achieved by modelling the conditional distribution for each pattern with a Gaussian Mixture Model for which the parameters are generated by a neural network. This thesis presents a novel method to introduce regularisation in this context for the special case where the mean and variance of the spherical Gaussian Kernels in the mixtures are fixed to predetermined values. Guidelines for how these parameters can be initialised are given, and it is shown how to apply the evidence framework to mixture density networks to achieve regularisation. This also provides an objective stopping criteria that can replace the `early stopping' methods that have previously been used. If the neural network used is an RBF network with fixed centres this opens up new opportunities for improved initialisation of the network weights, which are exploited to start training relatively close to the optimum. The new method is demonstrated on two data sets. The first is a simple synthetic data set while the second is a real life data set, namely satellite scatterometer data used to infer the wind speed and wind direction near the ocean surface. For both data sets the regularisation method performs well in comparison with earlier published results. Ideas on how the constraint on the kernels may be relaxed to allow fully adaptable kernels are presented.

Analysis and optimisation of the performance of nonlinear optical communication systems

We investigate the feasibility of simultaneous suppressing of the amplification noise and nonlinearity, representing the most fundamental limiting factors in modern optical communication. To accomplish this task we developed a general design optimisation technique, based on concepts of noise and nonlinearity management. We demonstrate the immense efficiency of the novel approach by applying it to a design optimisation of transmission lines with periodic dispersion compensation using Raman and hybrid Raman-EDFA amplification. Moreover, we showed, using nonlinearity management considerations, that the optimal performance in high bit-rate dispersion managed fibre systems with hybrid amplification is achieved for a certain amplifier spacing – which is different from commonly known optimal noise performance corresponding to fully distributed amplification. Required for an accurate estimation of the bit error rate, the complete knowledge of signal statistics is crucial for modern transmission links with strong inherent nonlinearity. Therefore, we implemented the advanced multicanonical Monte Carlo (MMC) method, acknowledged for its efficiency in estimating distribution tails. We have accurately computed acknowledged for its efficiency in estimating distribution tails. We have accurately computed marginal probability density functions for soliton parameters, by numerical modelling of Fokker-Plank equation applying the MMC simulation technique. Moreover, applying a powerful MMC method we have studied the BER penalty caused by deviations from the optimal decision level in systems employing in-line 2R optical regeneration. We have demonstrated that in such systems the analytical linear approximation that makes a better fit in the central part of the regenerator nonlinear transfer function produces more accurate approximation of the BER and BER penalty. We present a statistical analysis of RZ-DPSK optical signal at direct detection receiver with Mach-Zehnder interferometer demodulation

Seasonal variations of EPG levels in gastro-intestinal parasitic infection in a southeast asian controlled locale:a statistical analysis

We present a data based statistical study on the effects of seasonal variations in the growth rates of the gastro-intestinal (GI) parasitic infection in livestock. The alluded growth rate is estimated through the variation in the number of eggs per gram (EPG) of faeces in animals. In accordance with earlier studies, our analysis too shows that rainfall is the dominant variable in determining EPG infection rates compared to other macro-parameters like temperature and humidity. Our statistical analysis clearly indicates an oscillatory dependence of EPG levels on rainfall fluctuations. Monsoon recorded the highest infection with a comparative increase of at least 2.5 times compared to the next most infected period (summer). A least square fit of the EPG versus rainfall data indicates an approach towards a super diffusive (i. e. root mean square displacement growing faster than the square root of the elapsed time as obtained for simple diffusion) infection growth pattern regime for low rainfall regimes (technically defined as zeroth level dependence) that gets remarkably augmented for large rainfall zones. Our analysis further indicates that for low fluctuations in temperature (true on the bulk data), EPG level saturates beyond a critical value of the rainfall, a threshold that is expected to indicate the onset of the nonlinear regime. The probability density functions (PDFs) of the EPG data show oscillatory behavior in the large rainfall regime (greater than 500 mm), the frequency of oscillation, once again, being determined by the ambient wetness (rainfall, and humidity). Data recorded over three pilot projects spanning three measures of rainfall and humidity bear testimony to the universality of this statistical argument. © 2013 Chattopadhyay and Bandyopadhyay.

Rede Social Pessoal de Jovens Acolhidos em Lares de Infância e Juventude

A rede social dos sujeitos tem sido considerada fundamental para lidar com a adversidade. O objetivo central deste estudo foi caracterizar as redes sociais pessoais de jovens em regime de acolhimento institucional prolongado, na perspetiva dos próprios e dos técnicos das instituições e perceber se estas relacionam com o projeto de vida. Participaram 84 jovens, com idades compreendidas entre os 12 e os 20 anos de idade, acolhidos em 6 lares de infância e juventude do distrito de Santarém. Os participantes são na sua maioria do sexo feminino, com uma idade média de 15 anos. Para avaliação das variáveis em estudo utilizou-se a versão revista sumária do Instrumento de Avaliação de Redes Sociais Pessoais (IARSP-RS) e um questionário de autorresposta de caracterização da instituição de acolhimento. Os resultados desta investigação permitiram verificar que as redes dos jovens em situação de acolhimento institucional são fragmentadas, constituídas, em média, por 12 elementos e predominantemente compostas por familiares, amigos e profissionais das instituições, destacando-se as relações familiares. O nível de apoio social percebido é elevado na função emocional e informativa. Verificou-se uma maior reciprocidade do apoio por parte das jovens do sexo feminino, a institucionalização pode potenciar a expansão da rede social de suporte e a frequência de contactos associa-se à distância geográfica. A satisfação dos jovens com o suporte social é elevada. Metade dos jovens perspetivam a saída da instituição pela via da autonomização. A perceção dos profissionais das redes de suporte social é mais limitada, no tamanho, reciprocidade, satisfação com o suporte social e nas relações estabelecidas com a família, amigos da comunidade envolvente e relações de trabalho/estudo. Por outro lado, os técnicos identificam um nível mais elevado de densidade e de vínculos com os técnicos e educadores da instituição. Não se verificaram diferenças significativas nas características da rede social em função do projeto de vida, porém perceberam-se diferenças no projeto de vida em função da idade e da duração do acolhimento. A rede social apresenta um potencial protetor que vale a pena conhecer, fomentar e valorizar, destacando-se que é importante conhecer a perspetiva do próprio sujeito, já que esta tem um carácter muito pessoal. / The social network has been considered fundamental to deal with the adversities. The main objective of this study was to characterize the social network of the youngesters in prolonged institutional care, from the perspective of themselves and professionals of the institutions and to observe if their social network fits with their life project. 84 youngsters with ages between 12 and 20 years old living in 6 institutions of childhood and juvenile care from de district of Santarém participated in this study. The participants are mainly female, with an average age of 15 years old. To evaluate the variables in study, it was applied a version of the summary review of the Instrument of Evaluation of Personal Social Networks (IARSP-RS) and a survey of self-response to characterize the institutional care. The results of this investigation allowed to conclude that the social network of youngsters in a situation of institutional care are fragmented, composed by an average of 12 elements and primarily composed by family members, friends and professionals, but the most significant were the family relationships. It was observed that social support is high in terms of emotional and informative functions. It was observed an higher mutual support from the young ladies, institutionalization may foster the expansion of the social support network and that the frequency of connections is related with geographic distance. The satisfaction of the youngsters with social support is high. Half of the youngsters previews getting out of the institution by becoming autonomous. The understanding of the professionals of the social support network is more limited, about size, mutuality, satisfaction with the social support, family, friends of involving community and relationship of work/study. On the other side, the professionals identify an higher level of density and bonds with the technicals and educators of the institutions. Significative differences in the characteristics social network in terms of the life project weren’t found, however there were diferences in the life project in terms of the age and the duration of institutionalization. The social network has a protector potential that is worth of studying, promote and valorize, highlighting that it is important to know the perspective of the individual, because of its personal nature.

Videodensitometric analysis of advanced carotid plaque: correlation with MMP-9 and TIMP-1 expression

Background: Matrix metalloproteinase-9 (MMP-9) and tissue inhibitor of MMP (TIMP) promote derangement of the extracellular matrix, which is ultimately reflected in plaque images seen on ultrasound. Videodensitometry can identify structural disturbances in plaques. Objectives: To establish the correlations between values determined using videodensitometry in B-mode ultrasound images of advanced carotid plaques and the total expression of MMP-9 and TIMP-1 in these removed plaques. Methods: Thirty patients underwent ultrasonic tissue characterization of carotid plaques before surgery, using mean gray level (MGL), energy, entropy and homogeneity. Each patient was assigned preoperatively to one of 2 groups: group I, symptomatic patients (n = 16; 12 males; mean age 66.7 +/- 6.8 years), and group II, asymptomatic patients (n = 14; 8 males; mean age 67.6 +/- 6.81 years). Tissue specimens were analyzed for MMP-9 and TIMP-1 expression. Nine carotid arteries were used as normal tissue controls. Results: MMP-9 expression levels were elevated in group II and in normal tissues compared to group I (p < 0.001). TIMP-1 levels were higher in group II than in group I, and significantly higher in normal tissues than in group I (p = 0.039). The MGL was higher in group II compared to group I (p = 0.038). Energy had greater values in group II compared to group I (p = 0.02). There were no differences between patient groups in homogeneity and entropy. Energy positively correlated with MMP-9 and TIMP-1 expression (p = 0.012 and p = 0.031 respectively). Homogeneity positively correlated with MMP-9 and TIMP-1 expression (p = 0.034 and p = 0.047 respectively). There were no correlations between protein expression and MGL or entropy. Conclusions: Videodensitometric computer analysis of ultrasound scanning images can be used to identify stable carotid plaques, which have higher total expression levels of MMP-9 and TIMP-1 than unstable plaques.

Human gait recognition using extraction and fusion of global motion features

This paper proposes a novel computer vision approach that processes video sequences of people walking and then recognises those people by their gait. Human motion carries different information that can be analysed in various ways. The skeleton carries motion information about human joints, and the silhouette carries information about boundary motion of the human body. Moreover, binary and gray-level images contain different information about human movements. This work proposes to recover these different kinds of information to interpret the global motion of the human body based on four different segmented image models, using a fusion model to improve classification. Our proposed method considers the set of the segmented frames of each individual as a distinct class and each frame as an object of this class. The methodology applies background extraction using the Gaussian Mixture Model (GMM), a scale reduction based on the Wavelet Transform (WT) and feature extraction by Principal Component Analysis (PCA). We propose four new schemas for motion information capture: the Silhouette-Gray-Wavelet model (SGW) captures motion based on grey level variations; the Silhouette-Binary-Wavelet model (SBW) captures motion based on binary information; the Silhouette-Edge-Binary model (SEW) captures motion based on edge information and the Silhouette Skeleton Wavelet model (SSW) captures motion based on skeleton movement. The classification rates obtained separately from these four different models are then merged using a new proposed fusion technique. The results suggest excellent performance in terms of recognising people by their gait.

Phase-only optical encryption based on the zeroth-order phase-contrast technique

A phase-only encryption/decryption scheme with the readout based on the zeroth-order phase-contrast technique (ZOPCT), without the use of a phase-changing plate on the Fourier plane of an optical system based on the 4f optical correlator, is proposed. The encryption of a gray-level image is achieved by multiplying the phase distribution obtained directly from the gray-level image by a random phase distribution. The robustness of the encoding is assured by the nonlinearity intrinsic to the proposed phase-contrast method and the random phase distribution used in the encryption process. The experimental system has been implemented with liquid-crystal spatial modulators to generate phase-encrypted masks and a decrypting key. The advantage of this method is the easy scheme to recover the gray-level information from the decrypted phase-only mask applying the ZOPCT. An analysis of this decryption method was performed against brute force attacks. (C) 2009 Society of Photo-Optical Instrumentation Engineers. [DOI: 10.1117/1.3223629]

Brachial artery pulsatility index change 1 minute after 5-minute forearm compression - Comparison with flow-mediated dilatation

Objective. Endothelial impairment evaluation by sonographic measurement of flow-mediated dilatation (FMD) has become broadly used. However, this method has 2 main caveats: the dilatation depends on the baseline arterial diameter, and a high precision level is required. Vasodilatation leads to an amplified fall in impedance. We hypothesized that assessment of the pulsatility index change (PI-C) 1 minute after 5-minute forearm compression might evaluate that fall in impedance. The aim of this study was to compare the PI-C with FMD. Methods. Flow-mediated dilatation and the PI-C were assessed in 51 healthy women aged between 35.1 and 67.1 years. We correlated both FMD and the PI-C with age, body mass index, waist circumference, cholesterol level, high-density lipoprotein level, glucose level, systolic and diastolic blood pressure, pulse pressure, brachial artery diameter, simplified Framingham score, intima-media thickness, and carotid stiffness index. Intraclass correlation coefficients between 2 FMD and PI-C measurements were also examined. Results. Only FMD correlated with baseline brachial diameter (r=-0.53). The PI-C had a high correlation with age, body mass index, waist circumference, cholesterol level, systolic blood pressure, pulse pressure, simplified Framingham score, and intima-media thickness. The correlation between FMD and the PI-C was high (r=-0.66). The PI-C had a higher intraclass correlation coefficient (0.991) than FMD (0.836) but not brachial artery diameter (0.989). Conclusions. The PI-C had a large correlation with various markers of cardiovascular risk. Additionally, PI-C measurement does not require offline analysis, extra software, or electrocardiography We think that the PI-C could be considered a marker of endothelial function. However, more studies are required before further conclusions.

Detection and description of non-linear interdependence in normal multichannel human EEG data

Objectives: This study examines human scalp electroencephalographic (EEG) data for evidence of non-linear interdependence between posterior channels. The spectral and phase properties of those epochs of EEG exhibiting non-linear interdependence are studied. Methods: Scalp EEG data was collected from 40 healthy subjects. A technique for the detection of non-linear interdependence was applied to 2.048 s segments of posterior bipolar electrode data. Amplitude-adjusted phase-randomized surrogate data was used to statistically determine which EEG epochs exhibited non-linear interdependence. Results: Statistically significant evidence of non-linear interactions were evident in 2.9% (eyes open) to 4.8% (eyes closed) of the epochs. In the eyes-open recordings, these epochs exhibited a peak in the spectral and cross-spectral density functions at about 10 Hz. Two types of EEG epochs are evident in the eyes-closed recordings; one type exhibits a peak in the spectral density and cross-spectrum at 8 Hz. The other type has increased spectral and cross-spectral power across faster frequencies. Epochs identified as exhibiting non-linear interdependence display a tendency towards phase interdependencies across and between a broad range of frequencies. Conclusions: Non-linear interdependence is detectable in a small number of multichannel EEG epochs, and makes a contribution to the alpha rhythm. Non-linear interdependence produces spatially distributed activity that exhibits phase synchronization between oscillations present at different frequencies. The possible physiological significance of these findings are discussed with reference to the dynamical properties of neural systems and the role of synchronous activity in the neocortex. (C) 2002 Elsevier Science Ireland Ltd. All rights reserved.

Planeamento e optimização em redes celulares móveis auto-geridas assentes em LTE

O objectivo deste trabalho consiste em avaliar os benefícios das Self Organizing Networks (SON), no que concerne ao planeamento e optimização de redes Long Term Evolution (LTE), não só através do seu estudo, como também através do desenvolvimento e teste de algoritmos, que permitem avaliar o funcionamento de algumas das suas principais funções. O estudo efectuado sobre as SON permitiu identificar um conjunto de funções, tais como a atribuição automática de Physical Cell Id (PCI), o Automatic Neighbour Relation (ANR) e a optimização automática de parâmetros de handover, que permitem facilitar ou mesmo substituir algumas das tarefas mais comuns em planeamento e optimização de redes móveis celulares, em particular, redes LTE. Recorrendo a um simulador LTE destinado à investigação académica, em código aberto e desenvolvido em Matlab®, foi desenvolvido um conjunto de algoritmos que permitiram a implementação das funções em questão. Para além das funções implementadas, foram também introduzidas alterações que conferem a este simulador a capacidade de representar e simular redes reais, permitindo uma análise mais coerente dos algoritmos desenvolvidos. Os resultados obtidos, para além de evidenciarem claramente o benefício dos algoritmos desenvolvidos, foram ainda comparados com os obtidos pela ferramenta profissional de planeamento e optimização Atoll®, tendo-se verificado a franca proximidade de desempenho em algumas das funções. Finalmente, foi desenvolvida uma interface gráfica que permite o desenho, configuração e simulação de cenários, bem como a análise de resultados.

An Extension of Gompertzian Growth Dynamics Weibull and Frechet Models

In this work a new probabilistic and dynamical approach to an extension of the Gompertz law is proposed. A generalized family of probability density functions, designated by Beta* (p, q), which is proportional to the right hand side of the Tsoularis-Wallace model, is studied. In particular, for p = 2, the investigation is extended to the extreme value models of Weibull and Frechet type. These models, described by differential equations, are proportional to the hyper-Gompertz growth model. It is proved that the Beta* (2, q) densities are a power of betas mixture, and that its dynamics are determined by a non-linear coupling of probabilities. The dynamical analysis is performed using techniques of symbolic dynamics and the system complexity is measured using topological entropy. Generally, the natural history of a malignant tumour is reflected through bifurcation diagrams, in which are identified regions of regression, stability, bifurcation, chaos and terminus.

Self-Scheduling for energy and spinning reserve of wind/CSP plants by a MILP approach

This paper is on the self-scheduling for a power producer taking part in day-ahead joint energy and spinning reserve markets and aiming at a short-term coordination of wind power plants with concentrated solar power plants having thermal energy storage. The short-term coordination is formulated as a mixed-integer linear programming problem given as the maximization of profit subjected to technical operation constraints, including the ones related to a transmission line. Probability density functions are used to model the variability of the hourly wind speed and the solar irradiation in regard to a negative correlation. Case studies based on an Iberian Peninsula wind and concentrated solar power plants are presented, providing the optimal energy and spinning reserve for the short-term self-scheduling in order to unveil the coordination benefits and synergies between wind and solar resources. Results and sensitivity analysis are in favour of the coordination, showing an increase on profit, allowing for spinning reserve, reducing the need for curtailment, increasing the transmission line capacity factor. (C) 2014 Elsevier Ltd. All rights reserved.

Avaliação do potencial eólico para geração de energia eléctrica

Trabalho Final de Mestrado para a obtenção do grau de Mestre em Engenharia Mecânica /Energia

Measurement of the forward-backward asymmetry of electron and muon pair-production in pp collisions at s√ = 7 TeV with the ATLAS detector

This paper presents measurements from the ATLAS experiment of the forward-backward asymmetry in the reaction pp→Z/γ∗→l+l−, with l being electrons or muons, and the extraction of the effective weak mixing angle. The results are based on the full set of data collected in 2011 in pp collisions at the LHC at s√ = 7 TeV, corresponding to an integrated luminosity of 4.8 fb−1. The measured asymmetry values are found to be in agreement with the corresponding Standard Model predictions. The combination of the muon and electron channels yields a value of the effective weak mixing angle of 0.2308±0.0005(stat.)±0.0006(syst.)±0.0009(PDF), where the first uncertainty corresponds to data statistics, the second to systematic effects and the third to knowledge of the parton density functions. This result agrees with the current world average from the Particle Data Group fit.

Caracterização de uso e cobertura da terra na Amazônia utilizando imagens duais multitemporais do COSMO-SkyMed

A utilização de imagens de radar é fonte alternativa de informações para subsidiar o monitoramento da região amazônica, visto que as imagens ópticas têm limitações de imageamento em zonas tropicais face a ocorrência de nuvens. Por conseguinte este trabalho teve como objetivo analisar a capacidade das imagens-radar de banda X multitemporais e polarizadas obtidas pelo satélite COSMO-SkyMed (COnstellation of small Satellites for Mediterranean basin Observation), no modo intensidade, isoladamente e agregados às informações texturais, na caracterização temática de uso e cobertura da terra no município de Humaitá/AM. A metodologia empregada consistiu da: análise das imagens duais obtidas em duas aquisições subsequentes, de forma a explorar a potencialidade do conjunto de dados na forma quad-pol intensidade; extração dos atributos texturais a partir da matriz de coocorrência (Gray Level Co-occurrence Matrix) e posterior classificação contextual; avaliação estatística de desempenho temático das imagens intensidade e texturais, isoladas e em grupos polarizados. Dentre os vários resultados alcançados, foi verificado que o grupo formado somente pelas imagens intensidade apresentou o melhor desempenho, comparado àqueles contendo os atributos texturais. Nesta separabilidade, estavam envolvidas as classes de floresta, floresta aluvial, reflorestamento, savana, pasto e queimada, obtendo-se 66% de acurácia total e valor Kappa de 0,55. Os resultados mostraram que as imagens de banda X do COSMO-SkyMed, modo StripMap (Ping-Pong), multipolarizadas, têm potencial moderado para a caracterização e monitoramento da dinâmica de uso e cobertura da terra na Amazônia brasileira.