No description available.

No description available.

No description available.

Fourteen word frequency metrics were tested to evaluate their effectiveness in identifying vocabulary in a domain. Fifteen domain engineering projects were examined to measure how closely the vocabularies selected by the fourteen word frequency metrics were to the vocabularies produced by domain engineers. Six filtering mechanisms were also evaluated to measure their impact on selecting proper vocabulary terms. The results of the experiment show that stemming and stop word removal do improve overlap scores and that term frequency is a valuable contributor to overlap. Variations on term frequency are not always significant improvers of overlap. / Master of Science

Relation extraction is the problem of extracting relations between entities described in the text. Relations identify a common "fact" described by distinct entities. Conventional relation extraction approaches focus on supervised binary intra-sentence relations, where the assumption is relations only exist between two entities within the same sentence. These approaches have two key limitations. First, binary intra-sentence relation extraction methods can not extract a relation in a fact that is described by more than two entities. Second, these methods cannot extract relations that span more than one sentence, which commonly occurs as the number of entities increases. Third, these methods assume a supervised setting and are therefore not able to extract relations in the absence of sufficient labeled data for training. This work aims to overcome these limitations by developing n-ary cross-sentence relation extraction methods for both supervised and unsupervised settings. Our work has three main goals and contributions: (1) two unsupervised binary intra-sentence relation extraction methods, (2) a supervised n-ary cross-sentence relation extraction method, and (3) an unsupervised n-ary cross-sentence relation extraction method. To achieve these goals, our work includes the following contributions: (1) an automatic labeling method for n-ary cross-sentence data, which is essential for model training, (2) a reinforcement learning-based sentence distribution estimator to minimize the impact of noise on model training, (3) a generative clustering-based technique for intra-sentence unsupervised relation extraction, (4) a variational autoencoder-based technique for unsupervised n-ary cross-sentence relation extraction, and (5) a sentence group selector that identifies groups of sentences that form relations. / Master of Science / In this work, we designed multiple models to automatically extract relations from text. These relations represent the semantic connection between two or more proper nouns. Previous work includes models that can only extract relations between two proper nouns in a single sentence, while the methods proposed in this thesis can extract relations between two or more proper nouns in multiple sentences. We propose three models. The first model can automatically remove erroneous annotations in training data, thereby making the models more credible. We also propose a more effective model that can automatically extract relations between two proper nouns in a single sentence without the need for data annotation. We later extend this model so that it can extract relations between two or more proper nouns in multiple sentences.

Dans les opérations minières souterraines, la rentabilité et le niveau de sécurité dépendent également de la séquence d’extraction. Pour cela, cette dernière doit être constamment mise à jour, pour être adéquate et optimale. Ce mémoire présente les résultats de l’étude d’optimisation de la séquence d’extraction à la mine de Zinc de Perkoa, au Burkina Faso. Toutefois, l’étude s’est basée sur l’exploitation de la lentille principale, dans la Zone 310-400 du gisement. Au terme des travaux, une séquence pyramidale ascendante a été jugée favorable. Elle associera des chantiers transversaux et, longitudinaux primaires et secondaires, pour réduire le coût unitaire d’opération, tout en augmentant la flexibilité, ainsi que la productivité dans les chantiers d’abattage. Cependant, le sommet des pyramides et particulièrement, celui de la pyramide inversée dans la séquence, sera sous hautes contraintes induites. Par conséquent, un soutènement intensif du Toit des chantiers d’abattage et des épontes y est impératif. / Profitability and safety in underground mining operations also depend on stopes sequencing. Therefore, continuous update of a mining sequence makes it adequate and optimal. This thesis presents the results of a study carried out to optimize the extraction sequence at Perkoa Zinc Mine, located in Burkina Faso. The study is focused on mining of the portion located between level 310 and level 400 in the main lense of the orebody. This thesis proposes an ascendant pyramidal stoping sequence to mine this area. It also shows that a suitable combination of transverse and longitudinal primary and secondary stopes in the said sequence will reduce the operating cost, while increasing both flexibility and productivity. Nevertheless, accumulation of induced stresses will be observed on top of the pyramids. Particularly, the top of the inverse pyramids will be highly stressed. To enhance stability in such areas, a heavy ground support of backs and stopes walls is required.

A vast amount of usable electronic data is in the form of unstructured text. The relation extraction task aims to identify useful information in text (e.g., PersonW works for OrganisationX, GeneY encodes ProteinZ) and recode it in a format such as a relational database that can be more effectively used for querying and automated reasoning. However, adapting conventional relation extraction systems to new domains or tasks requires significant effort from annotators and developers. Furthermore, previous adaptation approaches based on bootstrapping start from example instances of the target relations, thus requiring that the correct relation type schema be known in advance. Generic relation extraction (GRE) addresses the adaptation problem by applying generic techniques that achieve comparable accuracy when transferred, without modification of model parameters, across domains and tasks. Previous work on GRE has relied extensively on various lexical and shallow syntactic indicators. I present new state-of-the-art models for GRE that incorporate governordependency information. I also introduce a dimensionality reduction step into the GRE relation characterisation sub-task, which serves to capture latent semantic information and leads to significant improvements over an unreduced model. Comparison of dimensionality reduction techniques suggests that latent Dirichlet allocation (LDA) – a probabilistic generative approach – successfully incorporates a larger and more interdependent feature set than a model based on singular value decomposition (SVD) and performs as well as or better than SVD on all experimental settings. Finally, I will introduce multi-document summarisation as an extrinsic test bed for GRE and present results which demonstrate that the relative performance of GRE models is consistent across tasks and that the GRE-based representation leads to significant improvements over a standard baseline from the literature. Taken together, the experimental results 1) show that GRE can be improved using dependency parsing and dimensionality reduction, 2) demonstrate the utility of GRE for the content selection step of extractive summarisation and 3) validate the GRE claim of modification-free adaptation for the first time with respect to both domain and task. This thesis also introduces data sets derived from publicly available corpora for the purpose of rigorous intrinsic evaluation in the news and biomedical domains.

Nous avons considéré autant les aspects thermodynamiques que cinétiques de la complexation d'ions métalliques (Cu2+ et Ni2+) par la 5-phénylazo-8-hydroxyquinoline (5Ph8HQ) et le cyclame greffés sur des nanoparticules de silice en dispersion colloïdale. Les silices pyrogénées, caractérisées par des surfaces spécifiques respectives d'environ 200 et 390 m2/g, ont été sélectionnées avec l'objectif d'obtenir des suspensions colloïdales stables. Nous avons démontré l'aptitude des colloïdes obtenus à extraire des cations divalents à l'état de traces (de l'ordre du micromolaire). Nous avons eu recours à l'ultrafiltration pour séparer la phase silicique dispersée de la phase aqueuse. Nous avons également montré l'intérêt de remplacer une pseudophase micellaire solubilisant l'extractant par nos phases solides. Nous avons décrit les cinétiques de complexation en utilisant la technique de la spectrophotométrie à écoulement bloqué pour les réactions les plus rapides. La dépendance de la constante de vitesse observée vis-à-vis de la concentration en cation métallique, du contre-ion considéré (acétate ou chlorure), de la force ionique et du pH a été évaluée. Pour décrire la cinétique de réaction du Ni2+ avec la 5Ph8HQ greffé, il nous a fallu prendre en compte les propriétés spectrophotométriques du complexe sur nanoparticules de silice. C'est un comportement original de la 5Ph8HQ sur silice puisque les densités optiques de solutions contenant la 5Ph8HQ en milieu micellaire ou le cyclame greffé sur silice suivent la loi de Beer-Lambert. L'étude du cyclame greffé qu'il s'agisse des cinétiques ou à l'équilibre est compliquée en raison du relargage de ce dernier en solution. / We considered both the thermodynamic and kinetic aspects of metal ions (Cu2+ and Ni2+) complexation by 5-phenylazo-8-hydroxyquinoline (5Ph8HQ) and cyclam grafted onto silica nanoparticles in colloidal dispersion. Fumed silicas with specific areas of respectively 200 and 390 m2/g were selected to get stable colloidal suspensions. We demonstrated the ability of these colloids to extract trace elements (at micromolar level). We used the ultrafiltration process to separate the dispersed silica phase from the aqueous phase. We also showed interest of our solid dispersed phases instead of a pseudo micellar one. We described complexation kinetics by stopped flow technique for the fastest reactions. The influence of metal cation concentration, counter-ion nature (acetate or chloride), ionic strength and pH on observed rate constant has been investigated. To describe the Ni2+ kinetics with grafted 5Ph8HQ, we took into account the spectrophotometric properties of the complex on silica nanoparticles. This is an original behaviour of 5Ph8HQ on silica since the optical densities of solutions containing 5Ph8HQ solubilised in micelles or cyclam grafted onto silica follow the Beer-Lambert's law. The study of grafted cyclam whether the kinetics or equilibrium is complicated due to the release of the latter in solution.

A research report submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science in Engineering (Metallurgy and Materials Engineering), May 2016 / The solvent extraction of iron and titanium from solution generated by the two-step sulphuric acid leaching of coal fly ash by Primene JMT was investigated. The influence of hydrogen ion concentration, Primene JMT concentration, aqueous to organic volume phase ratio and temperature on the extraction of iron and titanium was determined by the use of Design of Experiments. Hydrogen ion concentration and the interaction between the aqueous to organic volume phase ratio with Primene JMT concentration had a significant effect on the extraction of iron while temperature did not. Hydrogen ion concentration and temperature did not have a significant effect on the extraction of titanium, while the interaction between Primene JMT concentration and aqueous to organic volume phase ratio had a significant effect. Extraction improvement tests showed that at a hydrogen ion concentration of 0.28M, 88% iron and 99% extraction of titanium from coal-fly ash leach solution could be achieved. Construction of a McCabe-Thiele diagram showed that a four-stage solvent extraction system with Primene JMT could reduce the iron and titanium concentration in the coal fly ash leach solutions to below 0.05g/L / GS2016

This work aims to define the modes of action of a series of metal extraction ligands with particular focus on how these depend on the formation of supramolecular assemblies. Though solvent extraction processes are well established industrially often the understanding, particularly of the metal coordination chemistry, is less so. A greater understanding of a variety of solvent extraction systems can lead to the development of stronger and more specific extractants. Chapter 2 examines the role of inter-ligand interactions in the extraction of copper by phenolic oximes and pyrazoles. Computational methods are used to understand the importance of inter-ligand outer-sphere interactions in square-planar copper complexes. It is shown that functionalisation at different positions on the phenol ring can either stabilise or destabilise the copper complex and it is possible to predict the strength of extractants from DFT calculations. Substitution ortho to the phenolic oxygen in the oximes and pyrazoles can have a major effect of enhancing the strength of extractants by “buttressing” the H-bonding between ligands. However, in the amino-methyl substituted oximes buttressing is so strong that is has an adverse effect on complex formation. Crystal structures are confirmed by both ENDOR EPR spectroscopy and DFT structures. A series of 6-X-4-methyl-2-(5-alkyl-1H-pyrazol-3-yl)- phenols (X = H, OMe, Br and NO2) was synthesised and characterised (X = H, OMe, Br and NO2) and the copper extractant found to be 6-nitro-4-methyl-2-(5-(1,3,5-tri-methyl-pentyl)- 1H-pyrazol-3-yl)-phenol extractants. Computational DFT studies in the gas phase were carried out to calculate the formation energies of analogous phenolic pyrazole copper complexes. The predicted order of these energies followed the same trend shown by experimental solvent extraction studies. Studies also showed that substitution can affect not only complex stability through inter-ligand interactions through hydrogen bonding in the outer-sphere but also the strength of metal-ligand bonds. Chapter 3 looks at synergistic solvent extraction systems: where more than one extractant works together to provide additional strength and selectivity. Combinations of neutral N and O donor ligands with carboxylic, phosphinic and sulfonic acids were studied by solvent extraction, crystallographic and computational methods. Crystal structures and DFT-optimised structures show that ligands and acid form pseudo-tridentate ligands where both the neutral ligand and the deprotonated acid are coordinated directly to the metal centre with inter-ligand hydrogen bonding allowing for a more flexible backbone than a classic tridentate system. Although synergistic extractions systems often utilise carboxylic acids many of the structures show the similarities with systems containing phosphinic acids and it was shown experimentally that some extraction systems show greater synergism with phosphinic than carboxylic acid in the recovery of nickel. Chapter 4 deals with the extraction of molybdenum with commercial phosphinic acid extractant Cyanex 600. The propensity for molybdenum to form oxo clusters in aqueous solutions and the influence pH in both the speciation of the Mo species and extraction conditions contributes to a complex extraction profile. The pH dependence of extraction shows that different mechanisms operate at low (pH < 0) and high (pH > 0) pH. The extraction curve shows a conventional S-curve between pH 0 and 1.5 and slope analysis within this pH range gives a value very close to two but identification of structures which match this profile is complex. Maximum pH extraction is see at ~ pH 1.5. ESMS studies identified very similar species in the organic phase despite the variation seen in the S-curve. A survey of the structures of metal complexes of phosphinate ligands suggests that molybdenum-phosphinate complexes can often form cubane-like structure and negative ion ESMS data supports the concept of cluster formation in the organic phase. A common feature of the spectra are tetra- tri- and bi-metal- oxo species and spectra show a large number of peaks. It is very probable that the extraction of molybdenum(VI) with phosphinic acids is a dynamic system as extraction is influenced by the molybdenum speciation in the aqueous phase which is in turn influenced by both the pH and the molybdenum concentration both of which change over the course of a conventional extraction.