Factors influencing robustness and effectiveness of conditional random fields in active learning frameworks

Active learning approaches reduce the annotation cost required by traditional supervised approaches to reach the same effectiveness by actively selecting informative instances during the learning phase. However, effectiveness and robustness of the learnt models are influenced by a number of factors. In this paper we investigate the factors that affect the effectiveness, more specifically in terms of stability and robustness, of active learning models built using conditional random fields (CRFs) for information extraction applications. Stability, defined as a small variation of performance when small variation of the training data or a small variation of the parameters occur, is a major issue for machine learning models, but even more so in the active learning framework which aims to minimise the amount of training data required. The factors we investigate are a) the choice of incremental vs. standard active learning, b) the feature set used as a representation of the text (i.e., morphological features, syntactic features, or semantic features) and c) Gaussian prior variance as one of the important CRFs parameters. Our empirical findings show that incremental learning and the Gaussian prior variance lead to more stable and robust models across iterations. Our study also demonstrates that orthographical, morphological and contextual features as a group of basic features play an important role in learning effective models across all iterations.

Hydraulic conductivity fields: Gaussian or not?

Hydraulic conductivity (K) fields are used to parameterize groundwater flow and transport models. Numerical simulations require a detailed representation of the K field, synthesized to interpolate between available data. Several recent studies introduced high-resolution K data (HRK) at the Macro Dispersion Experiment (MADE) site, and used ground-penetrating radar (GPR) to delineate the main structural features of the aquifer. This paper describes a statistical analysis of these data, and the implications for K field modeling in alluvial aquifers. Two striking observations have emerged from this analysis. The first is that a simple fractional difference filter can have a profound effect on data histograms, organizing non-Gaussian ln K data into a coherent distribution. The second is that using GPR facies allows us to reproduce the significantly non-Gaussian shape seen in real HRK data profiles, using a simulated Gaussian ln K field in each facies. This illuminates a current controversy in the literature, between those who favor Gaussian ln K models, and those who observe non-Gaussian ln K fields. Both camps are correct, but at different scales.

Single step purification of plasmid DNA using peptide ligand affinity chromatography

Single step affinity chromatography was employed for the purification of plasmid DNA (pDNA), thus eliminating several steps compared with current commercial purification methods for pDNA. Significant reduction in pDNA production time and cost was obtained. This chromatographic operation employed a peptide-monolith construct to isolate pDNA from Escherichia coli (E. coli) impurities present in a clarified lysate feedstock. Mild conditions were applied to avoid any degradation of pDNA. The effect of some important parameters on pDNA yield was also evaluated with the aim of optimising the affinity purification of pDNA. The results demonstrate that 81% of pDNA was recovered and contaminating gDNA, RNA and protein were removed below detectable levels. © 2008 Elsevier B.V. All rights reserved.

Exploiting parallax in panoramic capture to construct light fields

We show that the parallax motion resulting from non-nodal rotation in panorama capture can be exploited for light field construction from commodity hardware. Automated panoramic image capture typically seeks to rotate a camera exactly about its nodal point, for which no parallax motion is observed. This can be difficult or impossible to achieve due to limitations of the mounting or optical systems, and consequently a wide range of captured panoramas suffer from parallax between images. We show that by capturing such imagery over a regular grid of camera poses, then appropriately transforming the captured imagery to a common parameterisation, a light field can be constructed. The resulting four-dimensional image encodes scene geometry as well as texture, allowing an increasingly rich range of light field processing techniques to be applied. Employing an Ocular Robotics REV25 camera pointing system, we demonstrate light field capture,refocusing and low-light image enhancement.

Different correlation of Sphingosine-1-Phosphate receptor 1 with receptor activator of nuclear factor kappa B ligand and regulatory T cells in rat periapical lesions

Introduction Sphingosine-1-phosphate receptor 1 (S1P1) is crucial for regulation of immunity and bone metabolism. This study aimed to investigate the expression of S1P1 in rat periapical lesions and its relationship with receptor activator of nuclear factor kappa B ligand (RANKL) and regulatory T (Treg) cells. Methods Periapical lesions were induced by pulp exposure in the first lower molars of 55 Wistar rats. Thirty rats were killed on days 0, 7, 14, 21, 28, and 35, and their mandibles were harvested for x-ray imaging, micro–computed tomography scanning, histologic observation, immunohistochemistry, enzyme histochemistry, and double immunofluorescence analysis. The remaining 25 rats were killed on days 0, 14, 21, 28, and 35, and mandibles were harvested for flow cytometry. Results The volume and area of the periapical lesions increased from day 0 to day 21 and then remained comparably stable after day 28. S1P1-positive cells were observed in the inflammatory periapical regions; the number of S1P1-positive cells peaked at day 14 and then decreased from day 21 to day 35. The distribution of S1P1-positive cells was positively correlated with the dynamics of RANKL-positive cells but was negatively correlated with that of Treg cells. Conclusions S1P1 expression was differentially correlated with RANKL and Treg cell infiltration in the periapical lesions and is therefore a contributing factor to the pathogenesis of such lesions.

The structures of the isomorphous potassium and rubidium salts of 4-nitrobenzoic acid and an overview of the metal complex stereochemistries of the alkali metal salt series with this ligand

The structures of the isomorphous potassium and rubidium polymeric coordination complexes with 4-nitrobenzoic acid, poly[mu2-aqua-aqua-mu3-(4-nitrobenzoato)-potassium], [K(C7H4N2O2)(H2O)2]n, (I) and poly[mu3-aqua-aqua-mu5-(4-nitrobenzoato)-rubidium], [Rb(C7H4N2O2)(H2O)2]n, (II) have been determined. In (I) the very distorted KO6 coordination sphere about the K+ centres in the repeat unit comprise two bridging nitro O-atom donors, a single bridging carboxyl O-atom donor and two water molecules, one of which is bridging. In the the Rb complex (II), the same basic MO6 coordination is found in the repeat unit but is expanded to RbO9 through a slight increase in the accepted Rb-O bond length range and includes an additional Rb-O(carboxyl) bond, completing a bidentate O,O'-chelate interaction, and additional bridging Rb-Onitro) and Rb-O(water) bonds. The comparative K-O and Rb-O bond length ranges are 2.738(3)-3.002(3)Ang. (I) and 2.884(2)-3.182(2)Ang. (II). The structure of (II) is also isomorphous as well as isostructural with the known structure of the nine-coordinate caesium 4-nitrobenzoate analogue, [Cs(C7H4N2O~2~)(H~2~O)2]n, (III) in which the Cs---O range is 3.047(4)-3.338(4)Ang. In all three complexes, common basic polymeric extensions are found, including two different centrosymmetric bridging interactions through both water and nitro groups as well as extensions along c through the p-related carboxyl group, giving a two-dimensional structure in (I). In (II) and (III), three-dimensional structures are generated through additional bridges through the nitro and water O-atoms. In all structures, both water molecules are involved in similar intra-polymer O-H...O hydrogen-bonding interactions to both carboxyl as well as water O-atom acceptors. A comparison of the varied coordination behaviour of the full set of Li-Cs salts with 4-nitrobenzoic acid is also made.

Aspect-based opinion mining from product reviews using conditional random fields

Product reviews are the foremost source of information for customers and manufacturers to help them make appropriate purchasing and production decisions. Natural language data is typically very sparse; the most common words are those that do not carry a lot of semantic content, and occurrences of any particular content-bearing word are rare, while co-occurrences of these words are rarer. Mining product aspects, along with corresponding opinions, is essential for Aspect-Based Opinion Mining (ABOM) as a result of the e-commerce revolution. Therefore, the need for automatic mining of reviews has reached a peak. In this work, we deal with ABOM as sequence labelling problem and propose a supervised extraction method to identify product aspects and corresponding opinions. We use Conditional Random Fields (CRFs) to solve the extraction problem and propose a feature function to enhance accuracy. The proposed method is evaluated using two different datasets. We also evaluate the effectiveness of feature function and the optimisation through multiple experiments.

Modelling of Electric Fields

South Africa has an electrical transmission grid of over 25 000 km of overhead power lines with voltages of 132 kV to 765 kV. The grid has been largely designed and built by the power utility, Eskom. This book embodies the planning philosophies, design principles and construction practices of Eskom. It is the culmination of decades of thought, study, research and the practical experience of many overhead power line engineers and researchers. The book covers the main aspects of overhead power line design and construction, from electrical first principles, system planning, insulation co-ordination (including live line working), mechanical design through to environmental impact management and power line communications. The content emphasises the need for close interaction between all technical disciplines involved and the importance of optimising designs for economy and performance. Additional challenges in South Africa are the relatively high altitude of the interior plateau (1 000 m to 1 700 m above sea level), severe lightning in some areas and long transmission distances. The book explains how these factors are accommodated in modern designs. Other advanced work covered includes the use and understanding of polymeric insulators, the judicious reduction of phase-to-phase spacings and the adoption of guyed structures.

Alteration of the C-terminal ligand specificity of the Erbin PDZ domain by allosteric mutational effects

Modulation of protein binding specificity is important for basic biology and for applied science. Here we explore how binding specificity is conveyed in PDZ (postsynaptic density protein-95/discs large/zonula occludens-1) domains, small interaction modules that recognize various proteins by binding to an extended C terminus. Our goal was to engineer variants of the Erbin PDZ domain with altered specificity for the most C-terminal position (position 0) where a Val is strongly preferred by the wild-type domain. We constructed a library of PDZ domains by randomizing residues in direct contact with position 0 and in a loop that is close to but does not contact position 0. We used phage display to select for PDZ variants that bind to 19 peptide ligands differing only at position 0. To verify that each obtained PDZ domain exhibited the correct binding specificity, we selected peptide ligands for each domain. Despite intensive efforts, we were only able to evolve Erbin PDZ domain variants with selectivity for the aliphatic C-terminal side chains Val, Ile and Leu. Interestingly, many PDZ domains with these three distinct specificities contained identical amino acids at positions that directly contact position 0 but differed in the loop that does not contact position 0. Computational modeling of the selected PDZ domains shows how slight conformational changes in the loop region propagate to the binding site and result in different binding specificities. Our results demonstrate that second-sphere residues could be crucial in determining protein binding specificity.

Behavior of simetryn and thiobencarb in the plough zone of rice fields

The behavior of simetryn and thiobencarb in flooded rice soil was investigated in a 2-year study. The concentrations of simetryn and thiobencarb were in the hundreds of μg kg^-1 in the top soil layer (0-5 cm) and became significantly lower in tens of μg kg^-1 in the deeper soil layers (5-10 and 10-15 cm). The half-lives of the two herbicides were also shorter (36 and 17 days for simetryn and thiobencarb, respectively) in the top soil layer, as they were most affected by environmental conditions, compared with corresponding values of 82 and 69 days in the 5-10 cm soil layer. Simetryn concentration was stable, while thiobencarb's half-life was 165 days in the 10-15 cm layer. About 35% of the applied mass of simetryn and thiobencarb were found in the rice soil compartment.

Export of radioactive cesium from agricultural fields under simulated rainfall in Fukushima

In this study, we investigated the impact of rainfall on runoff, soil erosion and consequently on the discharge of radioactive cesium in agricultural fields in Fukushima prefecture using a rainfall simulator. Simulated heavy rainfalls (50 mm h-1) generated significant runoff and soil erosion. The average concentration of radioactive cesium (the sum of 134Cs and 137Cs) in the runoff sediments was [similar]3500 Bq kg-1 dry soil, more than double the concentrations measured in the field soils which should be considered in studies using the 137Cs loss to estimate long-term soil erosion. However, the estimated mass of cesium discharged through one runoff event was less than 2% of the cesium inventory in the field. This suggested that cesium discharge via soil erosion is not a significant factor in reducing the radioactivity of contaminated soils in Fukushima prefecture. However, the eroded sediment carrying radioactive cesium will deposit into the river systems and potentially pose a radioactivity risk for aquatic living organisms.

Excess water storage depth - A water management practice to control simetryn and thiobencarb runoff from paddy fields

Experiments were carried out to verify the effectiveness of the excess water storage depth (EWSD) in reducing runoff losses of simetryn and thiobencarb from paddy fields upon appreciable rainfall events. A paddy plot having an EWSD of 2 cm was effective in controlling runoff with the herbicide losses of less than 1% of the applied herbicides. Meanwhile, a plot with 0-cm EWSD lost 18.1 and 3.7% of the applied mass of simetryn and thiobencarb, respectively. Therefore, an appropriate EWSD is essential during the recommended 7-day water holding period in order to completely hold the water inside the field in case of rainfall.

Applicability of ELISA in pesticide monitoring to control runoff of bensulfuron-methyl and simetryn from paddy fields

The applicability of ELISA kits was evaluated as an alternative to monitor bensulfuron-methyl and simetryn behavior in paddy water under intermittent (Plot 1) and continuous (Plot 2) irrigation schemes. Simetryn concentrations in both plots decreased exponentially from the peak of the first day. However, the simetryn kit systematically underestimated by a factor of 0.79 as compared to the GC method. Bensulfuron-methyl concentrations exhibited similar dissipation kinetics in paddy water and the drainage water. The bensulfuron-methyl kit was capable of distinguishing spatial variations of concentrations in the paddy field. The ELISA kits clearly indicated differences in the loss of both herbicides between the two plots and therefore may be useful for evaluating the water management practice of pesticide runoff control in paddy fields.

Behavior of pretilachlor and dimethametryn in water of flooded rice fields

In order to understand the behavior of pretilachlor, a popular rice herbicide in the world, and a synergetic active ingredient, dimethametryn, a monitoring study was conducted in 3 paddy plots in Kyushu region, Japan. The monitoring indicated different behaviors for both pesticides from those reported in the literature. Maximum concentrations of pretilachlor and dimethametryn were 1 order of magnitude lower than the values observed in previous studies. However, the dissipation rates estimated from monitoring data were in agreement with other studies in Japan. The pesticide product was tested and showed good dissolution of pretilachlor and dimethametryn in water, suggesting that another study is needed to explain the low concentrations of the two pesticides in the fields. Besides pesticide behaviors, it was observed from the monitoring that water management in paddy rice cultivation still requires more attention to reduce the environmental risk of rice pesticides.