Crustal growth in the 3.4-2.7Ga São José de Campestre Massif, Borborema Province, NE Brazil

The Archean (3.45-2.70Ga) rocks of the São José do Campestre Massif (SJCM) in the Borborema Province (NE Brazil) make up a small area (~6000km2) and are composed of granitoids and metasupracrustal rocks that define a complex magmatic and deformational history. The massif provides the opportunity to study mantle- and crustal-derived magmas generated since the Palaeoarchean. The orthogneisses of the SJCM are composed of: (1) tonalite to granodiorite with diorite enclaves (Bom Jesus gneiss, 3412±8Ma; TDM Nd model ages from 4.1 to 3.5Ga and negative epsilon Nd values); (2) biotite and ferroan-diopside monzogranite (Presidente Juscelino complex, 3356±21Ma and 3251±44Ma; TDM model ages range from 4.1 to 3.4Ga and epsilon Nd values that are slightly positive to negative); (3) hornblende tonalite to granodiorite (Brejinho complex, 3333±77Ma and 3187±8Ma; dominantly positive epsilon Nd values and TDM ages from 3.6 to 3.2Ga); (4) biotite monzogranite (São Pedro do Potengi gneiss, 3120±22Ma; TDM =3.5Ga; negative epsilon Nd value); (5) ferroan-diopside-grossular anorthosite and metagabbro (Senador Elói de Souza complex, 3033±3Ma); and (6) quartz diorite to syenogranite (São José do Campestre complex; 2685±9Ma and 2655±4Ma; negative epsilon Nd values and TDM ages from 3.9 to 3.3Ga). The orthogneisses are subalkaline to faintly alkaline, magnesian to ferroan, M- and I-type granitoids that follow either the K-enrichment or the trondhjemite trends. Each group has a subset with REE characteristics similar to Archean TTG and another that is analogous to Phanerozoic granitoids. They have negative Ta-Nb and Ti anomalies and have trace element contents of granitoids from subduction zones. Geochemical and Nd isotope data suggest that subducted oceanic crust and a depleted and metasomatised mantle wedge both acted as the magma sources. We propose a convergent tectonic model in which hybridisation of the upper mantle occurs through interactions with adakitic or trondhjemitic melts and recycling of earlier crust. The results imply that both the subducted oceanic crust and the mantle wedge played major roles in continent formation throughout successive episodes of arc accretion in Palaeo- and Mesoarchean times. The Archean rocks of the SJCM shares some similarities with the Pilbara, Kaapvaal, West African, and São Francisco cratons. However, the most reliable comparisons with the SJCM are with the neighbouring basement of the Nigeria and Cameroon shields. © 2012 Elsevier B.V.

Development of a CAD system for automatic classification of microcalcifications based on FPGA

This paper presents a Computer Aided Diagnosis (CAD) system that automatically classifies microcalcifications detected on digital mammograms into one of the five types proposed by Michele Le Gal, a classification scheme that allows radiologists to determine whether a breast tumor is malignant or not without the need for surgeries. The developed system uses a combination of wavelets and Artificial Neural Networks (ANN) and is executed on an Altera DE2-115 Development Kit, a kit containing a Field-Programmable Gate Array (FPGA) that allows the system to be smaller, cheaper and more energy efficient. Results have shown that the system was able to correctly classify 96.67% of test samples, which can be used as a second opinion by radiologists in breast cancer early diagnosis. (C) 2013 The Authors. Published by Elsevier B.V.

Mid-infrared supercontinuum generation in a suspended-core As2S3 chalcogenide microstructured optical fiber

We demonstrate the supercontinuum (SC) generation in a suspended-core As2S3 chalcogenide microstructured optical fiber (MOF). The variation of SC is investigated by changing the fiber length, pump peak power and pump wavelength. In the case of long fibers (20 and 40 cm), the SC ranges are discontinuous and stop at the wavelengths shorter than 3500 nm, due to the absorption of fiber. In the case of short fibers (1.3 and 2.4 cm), the SC ranges are continuous and can extend to the wavelengths longer than 4 μm. The SC broadening is observed when the pump peak power increases from 0.24 to 1.32 kW at 2500 nm. The SC range increases with the pump wavelength changing from 2200 to 2600 nm, corresponding to the dispersion of As2S3 MOF from the normal to anomalous region. The SC generation is simulated by the generalized nonlinear Schrödinger equation. The simulation includes the SC difference between 1.3 and 2.4 cm long fiber by 2500 nm pumping, the variation of SC with pump peak power in 2.4 cm long fiber, and the variation of SC with pump wavelength in 1.3 cm long fiber. The simulation agrees well with the experiment.

Exploring the virome of diseased horses

Metagenomics was used to characterize viral genomes in clinical specimens of horses with various organ-specific diseases of unknown aetiology. A novel parvovirus as well as a previously described hepacivirus closely related to human hepatitis C virus and equid herpesvirus 2 were identified in the cerebrospinal fluid of horses with neurological signs. Four co-infecting picobirnaviruses, including an unusual genome with fused RNA segments, and a divergent anellovirus were found in the plasma of two febrile horses. A novel cyclovirus genome was characterized from the nasal secretion of another febrile animal. Lastly, a small circular DNA genome with a Rep gene, from a virus we called kirkovirus, was identified in the liver and spleen of a horse with fatal idiopathic hepatopathy. This study expands the number of viruses found in horses, and characterizes their genomes to assist future epidemiological studies of their transmission and potential association with various equine diseases.

Active Learning to Maximize Area Under the ROC Curve

In active learning, a machine learning algorithmis given an unlabeled set of examples U, and is allowed to request labels for a relatively small subset of U to use for training. The goal is then to judiciously choose which examples in U to have labeled in order to optimize some performance criterion, e.g. classification accuracy. We study how active learning affects AUC. We examine two existing algorithms from the literature and present our own active learning algorithms designed to maximize the AUC of the hypothesis. One of our algorithms was consistently the top performer, and Closest Sampling from the literature often came in second behind it. When good posterior probability estimates were available, our heuristics were by far the best.

Bandit-Based Algorithms for Budgeted Learning

We explore the problem of budgeted machine learning, in which the learning algorithm has free access to the training examples’ labels but has to pay for each attribute that is specified. This learning model is appropriate in many areas, including medical applications. We present new algorithms for choosing which attributes to purchase of which examples in the budgeted learning model based on algorithms for the multi-armed bandit problem. All of our approaches outperformed the current state of the art. Furthermore, we present a new means for selecting an example to purchase after the attribute is selected, instead of selecting an example uniformly at random, which is typically done. Our new example selection method improved performance of all the algorithms we tested, both ours and those in the literature.

Balancing Exploration and Exploitation: A New Algorithm for Active Machine Learning

Active machine learning algorithms are used when large numbers of unlabeled examples are available and getting labels for them is costly (e.g. requiring consulting a human expert). Many conventional active learning algorithms focus on refining the decision boundary, at the expense of exploring new regions that the current hypothesis misclassifies. We propose a new active learning algorithm that balances such exploration with refining of the decision boundary by dynamically adjusting the probability to explore at each step. Our experimental results demonstrate improved performance on data sets that require extensive exploration while remaining competitive on data sets that do not. Our algorithm also shows significant tolerance of noise.

rho(0) photoproduction in AuAu collisions at root s(NN)=62.4 GeV measured with the STAR detector

Vector mesons may be photoproduced in relativistic heavy-ion collisions when a virtual photon emitted by one nucleus scatters from the other nucleus, emerging as a vector meson. The STAR Collaboration has previously presented measurements of coherent rho(0) photoproduction at center of mass energies of 130 GeV and 200 GeV in AuAu collisions. Here, we present a measurement of the cross section at 62.4 GeV; we find that the cross section for coherent rho(0) photoproduction with nuclear breakup is 10.5 +/- 1.5 +/- 1.6mb at 62.4 GeV. The cross-section ratio between 200 GeV and 62.4 GeV is 4.4 +/- 0.6, less than is predicted by most theoretical models. It is, however, proportionally much larger than the previously observed 15% +/- 55% increase between 130 GeV and 200 GeV.

Application of an active attachment model as a high-throughput demineralization biofilm model

Objectives: To investigate the potential of an active attachment biofilm model as a highthroughput demineralization biofilm model for the evaluation of caries-preventive agents. Methods: Streptococcus mutans UA159 biofilms were grown on bovine dentine discs in a highthroughput active attachment model. Biofilms were first formed in a medium with high buffer capacity for 24 h and then subjected to various photodynamic therapies (PACT) using the combination of Light Emitting Diodes (LEDs, Biotable (R)) and Photogem (R). Viability of the biofilms was evaluated by plate counts. To investigate treatment effects on dentine lesion formation, the treated biofilms were grown in a medium with low buffer capacity for an additional 24 h. Integrated mineral loss (IML) and lesion depth (LD) were assessed by transversal microradiography. Calcium release in the biofilm medium was measured by atomic absorption spectroscopy. Results: Compared to the water treated control group, significant reduction in viability of S. mutans biofilms was observed when the combination of LEDs and Photogem (R) was applied. LEDs or Photogem (R) only did not result in biofilm viability changes. Similar outcomes were also found for dentine lesion formation. Significant lower IML and LD values were only found in the group subjected to the combined treatment of LEDs and Photogem (R). There was a good correlation between the calcium release data and the IML or LD values. Conclusions: The high-throughput active attachment biofilm model is applicable for evaluating novel caries-preventive agents on both biofilm and demineralization inhibition. PACT had a killing effect on 24 h S. mutans biofilms and could inhibit the demineralization process. (C) 2011 Elsevier Ltd. All rights reserved.

System size and energy dependence of near-side dihadron correlations

Two-particle azimuthal (Delta phi) and pseudorapidity (Delta eta) correlations using a trigger particle with large transverse momentum (p(T)) in d+Au, Cu+Cu, and Au+Au collisions at root s(NN) = 62.4 GeV and 200 GeV from the STAR experiment at the Relativistic Heavy Ion Collider are presented. The near-side correlation is separated into a jet-like component, narrow in both Delta phi and Delta eta, and the ridge, narrow in Delta phi but broad in Delta eta. Both components are studied as a function of collision centrality, and the jet-like correlation is studied as a function of the trigger and associated p(T). The behavior of the jet-like component is remarkably consistent for different collision systems, suggesting it is produced by fragmentation. The width of the jet-like correlation is found to increase with the system size. The ridge, previously observed in Au+Au collisions at root s(NN) = 200 GeV, is also found in Cu+Cu collisions and in collisions at root s(NN) = 62.4 GeV, but is found to be substantially smaller at root s(NN) = 62.4 GeV than at root s(NN) = 200 GeV for the same average number of participants (< N-part >). Measurements of the ridge are compared to models.

Directed and elliptic flow of charged particles in Cu plus Cu collisions at root s(NN)=22.4 GeV

This paper reports results for directed flow v(1) and elliptic flow v(2) of charged particles in Cu + Cu collisions at root s(NN) = 22.4 GeV at the Relativistic Heavy Ion Collider. The measurements are for the 0-60% most central collisions, using charged particles observed in the STAR detector. Our measurements extend to 22.4-GeV Cu + Cu collisions the prior observation that v1 is independent of the system size at 62.4 and 200 GeV and also extend the scaling of v(1) with eta/y(beam) to this system. The measured v(2)(p(T)) in Cu + Cu collisions is similar for root s(NN) throughout the range 22.4 to 200 GeV. We also report a comparison with results from transport model (ultrarelativistic quantum molecular dynamics and multiphase transport model) calculations. The model results do not agree quantitatively with the measured v(1)(eta), v(2)(p(T)), and v(2)(eta).

Identified Hadron Compositions in p plus p and Au plus Au Collisions at High Transverse Momenta at root s(NN)=200 GeV

We report transverse momentum (p(T) <= 15 GeV/c) spectra of pi(+/-), K-+/-, p, (p) over bar, K-0(S), and rho(0) at midrapidity in p + p and Au + Au collisions at root s(NN) = 200 GeV. Perturbative QCD calculations are consistent with pi(+/-) spectra in p + p collisions but do not reproduce K and p((p) over bar) spectra. The observed decreasing antiparticle-to-particle ratios with increasing p(T) provide experimental evidence for varying quark and gluon jet contributions to high-p(T) hadron yields. The relative hadron abundances in Au + Au at p(T) >= 8 GeV/c are measured to be similar to the p + p results, despite the expected Casimir effect for parton energy loss.

Strangeness Enhancement in Cu-Cu and Au-Au Collisions at root s(NN)=200 GeV

We report new STAR measurements of midrapidity yields for the Lambda, (Lambda) over bar, K-S(0), Xi(-), (Xi) over bar (+), Omega(-), (Omega) over bar (+) particles in Cu + Cu collisions at root s(NN) = 200 GeV, and midrapidity yields for the Lambda, (Lambda) over bar, K-S(0) particles in Au + Au at root s(NN) = 200 GeV. We show that, at a given number of participating nucleons, the production of strange hadrons is higher in Cu + Cu collisions than in Au + Au collisions at the same center-of-mass energy. We find that aspects of the enhancement factors for all particles can be described by a parametrization based on the fraction of participants that undergo multiple collisions.

Anomalous centrality evolution of two-particle angular correlations from Au-Au collisions at root s(NN)=62 and 200 GeV

We present two-dimensional (2D) two-particle angular correlations measured with the STAR detector on relative pseudorapidity eta and azimuth phi for charged particles from Au-Au collisions at root s(NN) = 62 and 200 GeV with transverse momentum p(t) >= 0.15 GeV/c, vertical bar eta vertical bar <= 1, and 2 pi in azimuth. Observed correlations include a same-side (relative azimuth <pi/2) 2D peak, a closely related away-side azimuth dipole, and an azimuth quadrupole conventionally associated with elliptic flow. The same-side 2D peak and away-side dipole are explained by semihard parton scattering and fragmentation (minijets) in proton-proton and peripheral nucleus-nucleus collisions. Those structures follow N-N binary-collision scaling in Au-Au collisions until midcentrality, where a transition to a qualitatively different centrality trend occurs within one 10% centrality bin. Above the transition point the number of same-side and away-side correlated pairs increases rapidly relative to binary-collision scaling, the eta width of the same-side 2D peak also increases rapidly (eta elongation), and the phi width actually decreases significantly. Those centrality trends are in marked contrast with conventional expectations for jet quenching in a dense medium. The observed centrality trends are compared to perturbative QCD predictions computed in HIJING, which serve as a theoretical baseline, and to the expected trends for semihard parton scattering and fragmentation in a thermalized opaque medium predicted by theoretical calculations and phenomenological models. We are unable to reconcile a semihard parton scattering and fragmentation origin for the observed correlation structure and centrality trends with heavy-ion collision scenarios that invoke rapid parton thermalization. If the collision system turns out to be effectively opaque to few-GeV partons the present observations would be inconsistent with the minijet picture discussed here. DOI: 10.1103/PhysRevC.86.064902

Symmetry insights for design of supercomputer network topologies: roots and weights lattices

We present a family of networks whose local interconnection topologies are generated by the root vectors of a semi-simple complex Lie algebra. Cartan classification theorem of those algebras ensures those families of interconnection topologies to be exhaustive. The global arrangement of the network is defined in terms of integer or half-integer weight lattices. The mesh or torus topologies that network millions of processing cores, such as those in the IBM BlueGene series, are the simplest member of that category. The symmetries of the root systems of an algebra, manifested by their Weyl group, lends great convenience for the design and analysis of hardware architecture, algorithms and programs.