Towards a unified sentiment lexicon based on graphics processing units

This paper presents an approach to create what we have called a Unified Sentiment Lexicon (USL). This approach aims at aligning, unifying, and expanding the set of sentiment lexicons which are available on the web in order to increase their robustness of coverage. One problem related to the task of the automatic unification of different scores of sentiment lexicons is that there are multiple lexical entries for which the classification of positive, negative, or neutral {P, Z, N} depends on the unit of measurement used in the annotation methodology of the source sentiment lexicon. Our USL approach computes the unified strength of polarity of each lexical entry based on the Pearson correlation coefficient which measures how correlated lexical entries are with a value between 1 and -1, where 1 indicates that the lexical entries are perfectly correlated, 0 indicates no correlation, and -1 means they are perfectly inversely correlated and so is the UnifiedMetrics procedure for CPU and GPU, respectively. Another problem is the high processing time required for computing all the lexical entries in the unification task. Thus, the USL approach computes a subset of lexical entries in each of the 1344 GPU cores and uses parallel processing in order to unify 155802 lexical entries. The results of the analysis conducted using the USL approach show that the USL has 95.430 lexical entries, out of which there are 35.201 considered to be positive, 22.029 negative, and 38.200 neutral. Finally, the runtime was 10 minutes for 95.430 lexical entries; this allows a reduction of the time computing for the UnifiedMetrics by 3 times.

Parallel Computer Vision Algorithms for Graphics Processing Units

La evolución de los teléfonos móviles inteligentes, dotados de cámaras digitales, está provocando una creciente demanda de aplicaciones cada vez más complejas que necesitan algoritmos de visión artificial en tiempo real; puesto que el tamaño de las señales de vídeo no hace sino aumentar y en cambio el rendimiento de los procesadores de un solo núcleo se ha estancado, los nuevos algoritmos que se diseñen para visión artificial han de ser paralelos para poder ejecutarse en múltiples procesadores y ser computacionalmente escalables. Una de las clases de procesadores más interesantes en la actualidad se encuentra en las tarjetas gráficas (GPU), que son dispositivos que ofrecen un alto grado de paralelismo, un excelente rendimiento numérico y una creciente versatilidad, lo que los hace interesantes para llevar a cabo computación científica. En esta tesis se exploran dos aplicaciones de visión artificial que revisten una gran complejidad computacional y no pueden ser ejecutadas en tiempo real empleando procesadores tradicionales. En cambio, como se demuestra en esta tesis, la paralelización de las distintas subtareas y su implementación sobre una GPU arrojan los resultados deseados de ejecución con tasas de refresco interactivas. Asimismo, se propone una técnica para la evaluación rápida de funciones de complejidad arbitraria especialmente indicada para su uso en una GPU. En primer lugar se estudia la aplicación de técnicas de síntesis de imágenes virtuales a partir de únicamente dos cámaras lejanas y no paralelas—en contraste con la configuración habitual en TV 3D de cámaras cercanas y paralelas—con información de color y profundidad. Empleando filtros de mediana modificados para la elaboración de un mapa de profundidad virtual y proyecciones inversas, se comprueba que estas técnicas son adecuadas para una libre elección del punto de vista. Además, se demuestra que la codificación de la información de profundidad con respecto a un sistema de referencia global es sumamente perjudicial y debería ser evitada. Por otro lado se propone un sistema de detección de objetos móviles basado en técnicas de estimación de densidad con funciones locales. Este tipo de técnicas es muy adecuada para el modelado de escenas complejas con fondos multimodales, pero ha recibido poco uso debido a su gran complejidad computacional. El sistema propuesto, implementado en tiempo real sobre una GPU, incluye propuestas para la estimación dinámica de los anchos de banda de las funciones locales, actualización selectiva del modelo de fondo, actualización de la posición de las muestras de referencia del modelo de primer plano empleando un filtro de partículas multirregión y selección automática de regiones de interés para reducir el coste computacional. Los resultados, evaluados sobre diversas bases de datos y comparados con otros algoritmos del estado del arte, demuestran la gran versatilidad y calidad de la propuesta. Finalmente se propone un método para la aproximación de funciones arbitrarias empleando funciones continuas lineales a tramos, especialmente indicada para su implementación en una GPU mediante el uso de las unidades de filtraje de texturas, normalmente no utilizadas para cómputo numérico. La propuesta incluye un riguroso análisis matemático del error cometido en la aproximación en función del número de muestras empleadas, así como un método para la obtención de una partición cuasióptima del dominio de la función para minimizar el error. ABSTRACT The evolution of smartphones, all equipped with digital cameras, is driving a growing demand for ever more complex applications that need to rely on real-time computer vision algorithms. However, video signals are only increasing in size, whereas the performance of single-core processors has somewhat stagnated in the past few years. Consequently, new computer vision algorithms will need to be parallel to run on multiple processors and be computationally scalable. One of the most promising classes of processors nowadays can be found in graphics processing units (GPU). These are devices offering a high parallelism degree, excellent numerical performance and increasing versatility, which makes them interesting to run scientific computations. In this thesis, we explore two computer vision applications with a high computational complexity that precludes them from running in real time on traditional uniprocessors. However, we show that by parallelizing subtasks and implementing them on a GPU, both applications attain their goals of running at interactive frame rates. In addition, we propose a technique for fast evaluation of arbitrarily complex functions, specially designed for GPU implementation. First, we explore the application of depth-image–based rendering techniques to the unusual configuration of two convergent, wide baseline cameras, in contrast to the usual configuration used in 3D TV, which are narrow baseline, parallel cameras. By using a backward mapping approach with a depth inpainting scheme based on median filters, we show that these techniques are adequate for free viewpoint video applications. In addition, we show that referring depth information to a global reference system is ill-advised and should be avoided. Then, we propose a background subtraction system based on kernel density estimation techniques. These techniques are very adequate for modelling complex scenes featuring multimodal backgrounds, but have not been so popular due to their huge computational and memory complexity. The proposed system, implemented in real time on a GPU, features novel proposals for dynamic kernel bandwidth estimation for the background model, selective update of the background model, update of the position of reference samples of the foreground model using a multi-region particle filter, and automatic selection of regions of interest to reduce computational cost. The results, evaluated on several databases and compared to other state-of-the-art algorithms, demonstrate the high quality and versatility of our proposal. Finally, we propose a general method for the approximation of arbitrarily complex functions using continuous piecewise linear functions, specially formulated for GPU implementation by leveraging their texture filtering units, normally unused for numerical computation. Our proposal features a rigorous mathematical analysis of the approximation error in function of the number of samples, as well as a method to obtain a suboptimal partition of the domain of the function to minimize approximation error.

Molecular origin of the mosaic sequence arrangements of higher primate α-globin duplication units

The human adult α-globin locus consists of three pairs of homology blocks (X, Y, and Z) interspersed with three nonhomology blocks (I, II, and III), and three Alu family repeats, Alu1, Alu2, and Alu3. It has been suggested that an ancient primate α-globin-containing unit was ancestral to the X, Y, and Z and the Alu1/Alu2 repeats. However, the evolutionary origin of the three nonhomologous blocks has remained obscure. We have now analyzed the sequence organization of the entire adult α-globin locus of gibbon (Hylobates lar). DNA segments homologous to human block I occur in both duplication units of the gibbon α-globin locus. Detailed interspecies sequence comparisons suggest that nonhomologous blocks I and II, as well as another sequence, IV, were all part of the ancestral α-globin-containing unit prior to its tandem duplication. However, sometime thereafter, block I was deleted from the human α1-globin-containing unit, and block II was also deleted from the α2-globin-containing unit in both human and gibbon. These were probably independent events both mediated by independent illegitimate recombination processes. Interestingly, the end points of these deletions coincide with potential insertion sites of Alu family repeats. These results suggest that the shaping of DNA segments in eukaryotic genomes involved the retroposition of repetitive DNA elements in conjunction with simple DNA recombination processes.

Numbers and Organization of RNA Polymerases, Nascent Transcripts, and Transcription Units in HeLa Nuclei

Using HeLa cells, we have developed methods to determine 1) the number of RNA polymerases that are active at any moment, 2) the number of transcription sites, and 3) the number of polymerases associated with one transcription unit. To count engaged polymerases, cells were encapsulated in agarose, permeabilized, treated with ribonuclease, and the now-truncated transcripts extended in [32P]uridine triphosphate; then, the number of growing transcripts was calculated from the total number of nucleotides incorporated and the average increment in length of the transcripts. Approximately 15,000 transcripts were elongated by polymerase I, and ∼75,000 were elongated by polymerases II and III. Transcription sites were detected after the cells were grown in bromouridine for <2.5 min, after which the resulting bromo-RNA was labeled with gold particles; electron microscopy showed that most extranucleolar transcripts were concentrated in ∼2400 sites with diameters of ∼80 nm. The number of polymerases associated with a transcription unit was counted after templates were spread over a large area; most extranucleolar units were associated with one elongating complex. These results suggest that many templates are attached in a “cloud” of loops around a site; each site, or transcription “factory,” would contain ∼30 active polymerases and associated transcripts.

Annual league tables of mortality in neonatal intensive care units: longitudinal study

Objective: To assess whether crude league tables of mortality and league tables of risk adjusted mortality accurately reflect the performance of hospitals.

Randomised trial of educational visits to enhance use of systematic reviews in 25 obstetric units

Objective To evaluate the effectiveness of an educational visit to help obstetricians and midwives select and use evidence from a Cochrane database containing 600 systematic reviews.

Improved DNA binding specificity from polyzinc finger peptides by using strings of two-finger units

Multizinc finger peptides are likely to reach increased prominence in the search for the “ideal” designer transcription factor for in vivo applications such as gene therapy. However, for these treatments to be effective and safe, the peptides must bind with high affinity and, more importantly, with great specificity. Our previous research has shown that zinc finger arrays can be made to bind 18 bp of DNA with picomolar affinity, but also has suggested that arrays of fingers also may bind tightly to related sequences. This work addresses the question of zinc finger DNA binding specificity. We show that by changing the way in which zinc finger arrays are constructed—by linking three two-finger domains rather than two three-finger units—far greater target specificity can be achieved through increased discrimination against mutated or closely related sequences. These new peptides have the added capability of being able to span two short gaps of unbound DNA, although still binding with picomolar affinity to their target sites. We believe that this new method of constructing zinc finger arrays will offer greater efficacy in the fields of gene therapy and in the production of transgenic organisms than previously reported zinc finger arrays.

TectoRNA: modular assembly units for the construction of RNA nano-objects

Structural information on complex biological RNA molecules can be exploited to design tectoRNAs or artificial modular RNA units that can self-assemble through tertiary interactions thereby forming nanoscale RNA objects. The selective interactions of hairpin tetraloops with their receptors can be used to mediate tectoRNA assembly. Here we report on the modulation of the specificity and the strength of tectoRNA assembly (in the nanomolar to micromolar range) by variation of the length of the RNA subunits, the nature of their interacting motifs and the degree of flexibility of linker regions incorporated into the molecules. The association is also dependent on the concentration of magnesium. Monitoring of tectoRNA assembly by lead(II) cleavage protection indicates that some degree of structural flexibility is required for optimal binding. With tectoRNAs one can compare the binding affinities of different tertiary motifs and quantify the strength of individual interactions. Furthermore, in analogy to the synthons used in organic chemistry to synthesize more complex organic compounds, tectoRNAs form the basic assembly units for constructing complex RNA structures on the nanometer scale. Thus, tectoRNA provides a means for constructing molecular scaffoldings that organize functional modules in three-dimensional space for a wide range of applications.

Cyclic AMP Increases Cell Surface Expression of Functional Na,K-ATPase Units in Mammalian Cortical Collecting Duct Principal Cells

Cyclic AMP (cAMP) stimulates the transport of Na+ and Na,K-ATPase activity in the renal cortical collecting duct (CCD). The aim of this study was to investigate the mechanism whereby cAMP stimulates the Na,K-ATPase activity in microdissected rat CCDs and cultured mouse mpkCCDc14 collecting duct cells. db-cAMP (10−3 M) stimulated by 2-fold the activity of Na,K-ATPase from rat CCDs as well as the ouabain-sensitive component of 86Rb+ uptake by rat CCDs (1.7-fold) and cultured mouse CCD cells (1.5-fold). Pretreatment of rat CCDs with saponin increased the total Na,K-ATPase activity without further stimulation by db-cAMP. Western blotting performed after a biotinylation procedure revealed that db-cAMP increased the amount of Na,K-ATPase at the cell surface in both intact rat CCDs (1.7-fold) and cultured cells (1.3-fold), and that this increase was not related to changes in Na,K-ATPase internalization. Brefeldin A and low temperature (20°C) prevented both the db-cAMP-dependent increase in cell surface expression and activity of Na,K-ATPase in both intact rat CCDs and cultured cells. Pretreatment with the intracellular Ca2+ chelator bis-(o-aminophenoxy)-N,N,N′,N′-tetraacetic acid also blunted the increment in cell surface expression and activity of Na,K-ATPase caused by db-cAMP. In conclusion, these results strongly suggest that the cAMP-dependent stimulation of Na,K-ATPase activity in CCD results from the translocation of active pump units from an intracellular compartment to the plasma membrane.

Integration of complete transferred DNA units is dependent on the activity of virulence E2 protein of Agrobacterium tumefaciens.

Agrobacterium tumefaciens transfers transferred DNA (T-DNA), a single-stranded segment of its tumor-inducing (Ti) plasmid, to the plant cell nucleus. The Ti-plasmid-encoded virulence E2 (VirE2) protein expressed in the bacterium has single-stranded DNA (ssDNA)-binding properties and has been reported to act in the plant cell. This protein is thought to exert its influence on transfer efficiency by coating and accompanying the single-stranded T-DNA (ss-T-DNA) to the plant cell genome. Here, we analyze different putative roles of the VirE2 protein in the plant cell. In the absence of VirE2 protein, mainly truncated versions of the T-DNA are integrated. We infer that VirE2 protects the ss-T-DNA against nucleolytic attack during the transfer process and that it is interacting with the ss-T-DNA on its way to the plant cell nucleus. Furthermore, the VirE2 protein was found not to be involved in directing the ss-T-DNA to the plant cell nucleus in a manner dependent on a nuclear localization signal, a function which is carried by the NLS of VirD2. In addition, the efficiency of T-DNA integration into the plant genome was found to be VirE2 independent. We conclude that the VirE2 protein of A. tumefaciens is required to preserve the integrity of the T-DNA but does not contribute to the efficiency of the integration step per se.

Isolation of small polarized bile duct units.

Fragments of small interlobular bile ducts averaging 20 microns in diameter can be isolated from rat liver. These isolated bile duct units form luminal spaces that are impermeant to dextran-40 and expand in size when cultured in 10 microM forskolin for 24-48 hr. Secretion is Cl- and HCO3- dependent and is stimulated by forskolin > dibutyryl cAMP > secretion but not by dideoxyforskolin, as assessed by video imaging techniques. Secretin stimulates Cl-/HCO3- exchange activity, and intraluminal pH increases after forskolin administration. These studies establish that small polarized physiologically intact interlobular bile ducts can be isolated from rat liver. These isolated bile duct units should be useful preparations for assessing the transport properties of small bile duct segments, which are the primary site of injury in cholestatic liver disorders, known as "vanishing bile duct syndromes."

Recent landscape evolution at the "Barranco del Río Dulce Natural Park" (Spain). Landscape units and mapping

Landscape units based on the visual features of the relief have been distinguished in the “Barranco del Río Dulce Natural Park” (Spain). These units are geomorphic entities composed of several elementary landforms and characterized by a visual internal homogeneity, and contrast with other landscape units in their location, height, profile and gradients, reflecting their different evolution and genesis. Landscape units bear some subjectivity in their definition and in their boundary location due to the overlapping of geomorphic processes along time. Visual, compositional and conventional boundaries have been used for mapping. Neogene landscape evolution mainly occurred through thrust faulting at the Iberian Ranges-Tagus Basin boundary, driving tectonic uplift and erosion of the Ranges and correlative sedimentation in the Basin. Erosion of the Ranges occurred with the development of planation surfaces, leaving minor isolated reliefs in the upland plains landscape. The lowering of the base level, caused by the endorheic–exorheic transition of the Tagus Basin in the Pliocene, originates fluvial entrenchment and water table lowering with development of the first fluvial valleys and the capture of karstic depressions. Two subsequent phases of renewed fluvial incision (Pleistocene) lead to abandonment of some Pliocene valleys, fluvial captures, and development and reincision of tributaries