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.

Creació d'una eina en Java que a partir d'OWL i JESS generi SCORM

Aquest projecte consisteix en generar un programa escrit en llenguatge Java, que utilitzant un motor d'inferència a través d'una llibreria anomenada JESS, pugui llegir un document en format OWL que és una representació de l'ontologia (representació del coneixement) sobre una assignatura, transformant-lo al format de triples que és capaç d'interpretar JESS.

Adult consequences of late adolescent alcohol consumption: a systematic review of cohort studies.

Alcohol is responsible for a significant portion of the global burden of disease. There is widespread concern reported in the media and other sources about drinking trends among young people, particularly heavy episodic or “binge” drinking. Prominent among policy responses, in the UK and elsewhere, have been attempts to manage antisocial behaviour related to intoxication in public spaces. Much less attention has been given to the longer term effects of excessive drinking in adolescence on later adult health and well-being. Some studies suggest that individuals “mature out” of late adolescent drinking behaviour, whilst others identify enduring effects on drinking and broader health and social outcomes in adulthood. If adolescent drinking does not cause later difficulties in adulthood then intervention approaches aimed at addressing the acute consequences of alcohol, such as unintentional injuries and anti-social behaviour, may be the most appropriate solution. If causal relationships do exist, however, this approach will not address the cumulative harms produced by alcohol, unless such intervention successfully modifies the long-term relationship with alcohol, which seems unlikely. To address this issue a systematic review of cohort studies was conducted, as this approach provides the strongest observational study design to evaluate evidence for causal inference.This resource was contributed by The National Documentation Centre on Drug Use.

Forensic science - A true science ?

While the US jurisprudence of the 1993 Daubert requires judges to question not only the methodology behind, but also the principles governing, a body of knowledge to qualify it as scientific, can forensic science, based on Locard's and Kirk's Principles, pretend to this higher status in the courtroom ? Moving away from the disputable American legal debate, this historical and philosophical study will screen the relevance of the different logical epistemologies to recognize the scientific status of forensic science. As a consequence, the authors are supporting a call for its recognition as a science of its own, defined as the science of identifying and associating traces for investigative and security purposes, based o its fundamental principles and the case assesment and interpretation process that follows with its specific and relevant mode of inference.

Naturally acquired antibodies to merozoite surface protein (MSP)-1(19) and cumulative exposure to Plasmodium falciparum and Plasmodium vivax in remote populations of the Amazon Basin of Brazil

To infer recent patterns of malaria transmission, we measured naturally acquired IgG antibodies to the conserved 19-kDa C-terminal region of the merozoite surface protein (MSP)-1 of both Plasmodium vivax (PvMSP-1(19)) and Plasmodium falciparum (PfMSP-1(19)) in remote malaria-exposed populations of the Amazon Basin. Community-based cross-sectional surveys were carried out between 2002 and 2003 in subjects of all age groups living along the margins of the Unini and Jaú rivers, Northwestern Brazil. We found high prevalence rates of IgG antibodies to PvMSP-1(19) (64.0 - 69.6%) and PfMSP-1(19) (51.6 - 52.0%), with significant differences in the proportion of subjects with antibodies to PvMSP-1(19) according to age, place of residence and habitual involvement in high-risk activities, defining some groups of highly exposed people who might be preferential targets of malaria control measures. In contrast, no risk factor other than age was significantly associated with seropositivity to PfMSP-1(19). Only 14.1% and 19.3% of the subjects tested for antibodies to PvMSP-1(19) and PfMSP-1(19) in consecutive surveys (142 - 203 days apart) seroconverted or had a three fold or higher increase in the levels of antibodies to these antigens. We discuss the extent to which serological data correlated with the classical malariometric indices and morbidity indicators measured in the studied population at the time of the seroprevalence surveys and highlight some limitations of serological data for epidemiological inference.

Total hemoglobin mass--a new parameter to detect blood doping?

PURPOSE: All kinds of blood manipulations aim to increase the total hemoglobin mass (tHb-mass). To establish tHb-mass as an effective screening parameter for detecting blood doping, the knowledge of its normal variation over time is necessary. The aim of the present study, therefore, was to determine the intraindividual variance of tHb-mass in elite athletes during a training year emphasizing off, training, and race seasons at sea level. METHODS: tHb-mass and hemoglobin concentration ([Hb]) were determined in 24 endurance athletes five times during a year and were compared with a control group (n = 6). An analysis of covariance was used to test the effects of training phases, age, gender, competition level, body mass, and training volume. Three error models, based on 1) a total percentage error of measurement, 2) the combination of a typical percentage error (TE) of analytical origin with an absolute SD of biological origin, and 3) between-subject and within-subject variance components as obtained by an analysis of variance, were tested. RESULTS: In addition to the expected influence of performance status, the main results were that the effects of training volume (P = 0.20) and training phases (P = 0.81) on tHb-mass were not significant. We found that within-subject variations mainly have an analytical origin (TE approximately 1.4%) and a very small SD (7.5 g) of biological origin. CONCLUSION: tHb-mass shows very low individual oscillations during a training year (<6%), and these oscillations are below the expected changes in tHb-mass due to Herythropoetin (EPO) application or blood infusion (approximately 10%). The high stability of tHb-mass over a period of 1 year suggests that it should be included in an athlete's biological passport and analyzed by recently developed probabilistic inference techniques that define subject-based reference ranges.

Polymorphism analysis of the CTLA-4 gene in paracoccidioidomycosis patients

The CTLA-4 protein is expressed in activated T cells and plays an essential role in the immune response through its regulatory effect on T cell activation. Polymorphisms of the CTLA-4 gene have been correlated with autoimmune, neoplastic and infectious illnesses. This work aimed to verify possible associations between single nucleotide polymorphisms (SNPs) in CTLA-4, -318C/T in the promoter and +49A/G in exon 1 and paracoccidioidomycosis (PCM) caused by Paracoccidioides brasiliensis. For this purpose, 66 chronic form PCM patients and 76 healthy controls had their allele, genotype and haplotype frequencies determined. The genetic admixture structure of the patients and controls was evaluated to eliminate ancestral bias. The comparison of frequencies indicated no significant differences between patients and controls that could link the SNPs to PCM. Groups were admixture matched with no difference observed in population ancestry inference, indicating that the absence of association between CTLA-4 polymorphisms and PCM could not be attributed to ancestral bias. This study showed that there was no association between the CTLA-4 SNPs -318 and +49 and the resistance or susceptibility to PCM.

Evolution of asexuality via different mechanisms in grass thrips (thysanoptera: Aptinothrips).

Asexual lineages can derive from sexual ancestors via different mechanisms and at variable rates, which affects the diversity of the asexual population and thereby its ecological success. We investigated the variation and evolution of reproductive systems in Aptinothrips, a genus of grass thrips comprising four species. Extensive population surveys and breeding experiments indicated sexual reproduction in A. elegans, asexuality in A. stylifer and A. karnyi, and both sexual and asexual lineages in A. rufus. Asexuality in A. stylifer and A. rufus coincides with a worldwide distribution, with sexual A. rufus lineages confined to a limited area. Inference of molecular phylogenies and antibiotic treatment revealed different causes of asexuality in different species. Asexuality in A. stylifer and A. karnyi has most likely genetic causes, while it is induced by endosymbionts in A. rufus. Endosymbiont-community characterization revealed presence of Wolbachia, and lack of other bacteria known to manipulate host reproduction. However, only 69% asexual A. rufus females are Wolbachia-infected, indicating that either an undescribed endosymbiont causes asexuality in this species or that Wolbachia was lost in several lineages that remained asexual. These results open new perspectives for studies on the maintenance of mixed sexual and asexual reproduction in natural populations.

Redes neuronales multimodelo aplicadas al control de sistemas

En este proyecto se plantea un nuevo modelo de red neuronal llamado multimodelo. Si en un modelo convencional de red neuronal obtenemos la salida a partir de una serie de parámetros, en un sistema multimodelo recalculamos la salida a partir de los mismos valores de entrada, pero variando los parámetros. Con esto se consigue una mayor muestra de salidas, que permiten decidir cuales son los parámetros más acertados (los que devuelven una salida más cercana a la salida deseada del sistema de control). Para llevar a cabo este proyecto se ha hecho uso de Matlab 7.3 (R2006b).

The hypogene iron oxide copper-gold mineralization in the Mantoverde District, Northern Chile

The Mantoverde iron oxide copper-gold (IOCC) district, northern Chile, is known for its Cu production from supergene ores. Recently, exploration outlined an additional hypogene ore resource of 440 Mt with 0.56 percent Cu, and 0.12 g/t An. The hypogene sulfide mineralization occurs mainly as chalcopyrite and pyrite, typically in specularite or magnetite-cemented breccias and associated stockworks. The host rocks underwent variably intense K feldspar alteration, chloritization, sericitization, silicification, and/or carbonatization. A district scale Na(-Ca) alteration is absent. The IOCC mineralization in the district shows a strong tectonic control by northwest- to north-northwest-trending brittle structures. Large Cu sulfide-rich veins or Cu sulfide-cemented breccias are absent. Therefore, head grades of 4 percent Cu are an exception. There is a positive correlation between Cu and An grades. Gold is probably contained mostly in chalcopyrite and pyrite. Elevated concentrations of light rare-earth elements (LREE) occur locally but are attributed to redistribution of LREE within the deposits rather than to derivation from external sources. The Cu-Au ores in the Mantoverde district are low in and have relatively low contents in heavy metals that are potentially hazardous to the environment, such as As (avg 14 ppm), Hg (<5 ppm), or Cd (<0.2 ppm). The sulfur isotope ratios of chalcopyrite from the IOCC deposits lie between -5.6 and 8.9 per mil delta(34)S(VCDT). They show systematic variations within the district, which are interpreted to reflect relative distance to inferred fluid conduits and the level of deposition within the hydrothermal system. Most initial (87)Sr/(86)Sr values of altered volcanic rocks and hydrothermal calcite from the Mantoverde district are between 0.7031 and 0.7060 and are similar to those of the igneous rocks of the region. Lead isotope ratios of chalcopyrite are consistent with Pb (and by inference Cu) derived from Early Cretaceous magmatism. The sulfur, strontium, and lead isotope data of chalcopyrite, calcite gangue, or altered host rocks, respectively, are compatible with a genetic model that involves cooling of metal and sulfur-bearing magmatic-hydrothermal fluids that mix with meteoric waters or seawater at relatively shallow crustal levels. An additional exotic sulfur input is likely, though not required, for the copper mineralization. Apart from the IOCC. deposits, there are a number of smaller magnetite(-apatite) bodies in the district. These are geologically similar to the Cu-Au-bearing magnetite bodies, but are related to splays of the north-south-trending Atacama fault zone and differ in alteration and texture.

Uncertainty Quantification Of A Semi-Supervised Support Vector Regression Reservoir Model

Uncertainty quantification of petroleum reservoir models is one of the present challenges, which is usually approached with a wide range of geostatistical tools linked with statistical optimisation or/and inference algorithms. Recent advances in machine learning offer a novel approach to model spatial distribution of petrophysical properties in complex reservoirs alternative to geostatistics. The approach is based of semisupervised learning, which handles both ?labelled? observed data and ?unlabelled? data, which have no measured value but describe prior knowledge and other relevant data in forms of manifolds in the input space where the modelled property is continuous. Proposed semi-supervised Support Vector Regression (SVR) model has demonstrated its capability to represent realistic geological features and describe stochastic variability and non-uniqueness of spatial properties. On the other hand, it is able to capture and preserve key spatial dependencies such as connectivity of high permeability geo-bodies, which is often difficult in contemporary petroleum reservoir studies. Semi-supervised SVR as a data driven algorithm is designed to integrate various kind of conditioning information and learn dependences from it. The semi-supervised SVR model is able to balance signal/noise levels and control the prior belief in available data. In this work, stochastic semi-supervised SVR geomodel is integrated into Bayesian framework to quantify uncertainty of reservoir production with multiple models fitted to past dynamic observations (production history). Multiple history matched models are obtained using stochastic sampling and/or MCMC-based inference algorithms, which evaluate posterior probability distribution. Uncertainty of the model is described by posterior probability of the model parameters that represent key geological properties: spatial correlation size, continuity strength, smoothness/variability of spatial property distribution. The developed approach is illustrated with a fluvial reservoir case. The resulting probabilistic production forecasts are described by uncertainty envelopes. The paper compares the performance of the models with different combinations of unknown parameters and discusses sensitivity issues.

Evolutionary simulations to detect functional lineage-specific genes.

MOTIVATION: Supporting the functionality of recent duplicate gene copies is usually difficult, owing to high sequence similarity between duplicate counterparts and shallow phylogenies, which hamper both the statistical and experimental inference. RESULTS: We developed an integrated evolutionary approach to identify functional duplicate gene copies and other lineage-specific genes. By repeatedly simulating neutral evolution, our method estimates the probability that an ORF was selectively conserved and is therefore likely to represent a bona fide coding region. In parallel, our method tests whether the accumulation of non-synonymous substitutions reveals signatures of selective constraint. We show that our approach has high power to identify functional lineage-specific genes using simulated and real data. For example, a coding region of average length (approximately 1400 bp), restricted to hominoids, can be predicted to be functional in approximately 94-100% of cases. Notably, the method may support functionality for instances where classical selection tests based on the ratio of non-synonymous to synonymous substitutions fail to reveal signatures of selection. Our method is available as an automated tool, ReEVOLVER, which will also be useful to systematically detect functional lineage-specific genes of closely related species on a large scale. AVAILABILITY: ReEVOLVER is available at http://www.unil.ch/cig/page7858.html.

Turnover and accumulation of genetic diversity across large time-scale cycles of isolation and connection of populations.

Major climatic and geological events but also population history (secondary contacts) have generated cycles of population isolation and connection of long and short periods. Recent empirical and theoretical studies suggest that fast evolutionary processes might be triggered by such events, as commonly illustrated in ecology by the adaptive radiation of cichlid fishes (isolation and reconnection of lakes and watersheds) and in epidemiology by the fast adaptation of the influenza virus (isolation and reconnection in hosts). We test whether cyclic population isolation and connection provide the raw material (standing genetic variation) for species evolution and diversification. Our analytical results demonstrate that population isolation and connection can provide, to populations, a high excess of genetic diversity compared with what is expected at equilibrium. This excess is either cyclic (high allele turnover) or cumulates with time depending on the duration of the isolation and the connection periods and the mutation rate. We show that diversification rates of animal clades are associated with specific periods of climatic cycles in the Quaternary. We finally discuss the importance of our results for macroevolutionary patterns and for the inference of population history from genomic data.