No description available.

No description available.

No description available.

Annotated text corpora facilitate the linguistic investigation of language as well as the automation of natural language processing (NLP) tasks. NLP tasks include problems such as spam email detection, grammatical analysis, and identifying mentions of people, places, and events in text. However, constructing high quality annotated corpora can be expensive. Cost can be reduced by employing low-cost internet workers in a practice known as crowdsourcing, but the resulting annotations are often inaccurate, decreasing the usefulness of a corpus. This inaccuracy is typically mitigated by collecting multiple redundant judgments and aggregating them (e.g., via majority vote) to produce high quality consensus answers. We improve the quality of consensus labels inferred from imperfect annotations in a number of ways. We show that transfer learning can be used to derive benefit from out-dated annotations which would typically be discarded. We show that, contrary to popular preference, annotation aggregation models that take a generative data modeling approach tend to outperform those that take a condition approach. We leverage this insight to develop csLDA, a novel annotation aggregation model that improves on the state of the art for a variety of annotation tasks. When data does not permit generative data modeling, we identify a conditional data modeling approach based on vector-space text representations that achieves state-of-the-art results on several unusual semantic annotation tasks. Finally, we identify a family of models capable of aggregating annotation data containing heterogenous annotation types such as label frequencies and labeled features. We present a multiannotator active learning algorithm for this model family that jointly selects an annotator, data items, and annotation type.

Understanding thickness variation and facies transitions in the mudstone-rich part of the Upper Middle Jurassic (Callovian) Entrada Sandstone depositional system is critical for constraining the paleogeography and evaluating the economic potential of Utah's Entrada Sandstone. Facies of the Entrada Sandstone in south-central Utah are dominated by mudstone-rich intertidal facies that were widespread within the Jurassic seaway. Intertidal deposits interfinger basinward with subtidal ooid-bearing shoals and bars, and landward supratidal sabkha, and erg-margin eolian deposits. Three sections were measured to improve understanding of the lateral and vertical facies transitions. Variations in thickness indicate the rate of developing accommodation space was high along the southwestern shoreline and relatively low along the northeastern shoreline during Callovian time. Although accommodation space was highest in the west, sediment supply from the west kept pace with, and eventually outpaced subsidence. In the east, sediment supply was significant but at one time was outpaced by subsidence, creating a complete range of facies, from subtidal to supratidal deposits. Along this eastern shoreline, erg-margin coastal dunes associated with the larger erg to the east eventually prograded westward. The variation in subsidence, sediment supply, and sediment source makes sequence stratigraphic correlation difficult. Reservoir-quality sandstones are associated with muddy sections of the Entrada Sandstone within the San Rafael Swell. Porosity and permeability of the facies in this area indicate excellent reservoir potential in three of eight facies that were studied. Porosities of these potential reservoirs ranged from 11-22%, with permeabilities ranging from 44-430md. These high quality reservoir facies are surrounded by muddy, low reservoir-quality rocks, creating conditions amenable to the development of stratigraphic hydrocarbon traps. Based on further study and a modern analog at the north of the Gulf of California, Hicks and others' (2010) depositional model for the Entrada Sandstone of south-central Utah has been modified to include newly interpreted facies. This improved depositional model may have predictive power in exploring for stratigraphic and combination traps within the Entrada system of Utah and analogous depositional systems throughout the world.

<p> An expression is obtained for the energy per particle in neutron star matter. The energy per particle is expressed as a function of, y, the ratio of protons to the total number of nucleons in the system. Minimizing the energy with respect to y gives the optimum proton ratio at a given density. Using an effective nuclear force, the results were extrapolated to a density of p = 6pNM. The proton ratio is rather sensitive to the force used, but all forces used indicated a peak in the proton concentration at p (approximately equal to) 2pNM. The expression for the energy as a function of y was also used to interpolate the energy per particle between the nuclear matter and neutron . gas limits. The form of this interpolation is important in determining the stability of neutron-rich nuclei. </p> / Thesis / Master of Science (MSc)

<p> Sulfide-poor platinum-group element (PGE) mineralization occurs within the Archean Boston Creek Flow ferropicrite, Abitibi greenstone belt, Canada. The PGE mineralization (∑PGE+Au = up to 1000 ppb) is manifest as metre-scale platinum-group mineral-bearing pods of disseminated chalcopyrite and pyrite (< 1 modal%) within titaniferous magnetite-rich gabbroic rock at the base of its central gabbroic layer. This mineralization is distinct in character from well known PGE mineralization associated with massive Fe-Ni-Cu sulfides at the base of komatiite flows at Kambalda, Western Australia and elsewhere. Exploration strategies presently used to search for PGE in mafic and ultramafic volcanic rock terrains are based on the geological and geophysical characteristics of sulfide-rich PGE mineralization. Consequently, refinements in exploration strategies are required if economic concentrations of sulfide-poor PGE mineralization are to be discovered in volcanic terrains.</p> <p> To begin development of such exploration criteria, ground-based magnetic and VLF surveys were conducted over the PGE mineralization along a single cross-section through the BCF. Drill core samples were collected along this transect to characterize the volume magnetic susceptibility and natural remanent magnetization (NRM) of the mineralization. Magnetic highs ranging in intensity from 64000 to 65000 nT were recorded for the base of the gabbroic layer, including mineralized outcrops. Susceptibilities of up to 9700 cgs and high remanence values with variable directions were determined. VLF, as expected, was not useful in identifying the mineralized horizons. The peridotite at the base of the flow appears to be the only conductive rock in the BCF.</p> <p> The magnetic highs associated with the base of the gabbroic layer define a positive anomaly that appears to be podiform in outline and up to ten metres in maximum dimension. This result suggests that the titaniferous magnetite-rich rock is itself podiform, like the enclosed PGE mineralization. This magnetic anomaly is not extensive enough to be evident on a regional scale aeromagnetic map as a separate anomaly within the BCF, and is not evident through VLF techniques. The rocks hosting the PGE mineralization is defined by the paleomagnetic results but this is not a practical field method.</p> <p> The geophysical characterization of the PGE mineralization host rocks, and of the BCF in general, demonstrates the potential of detailed magnetic and susceptibility mapping, together with petrographic and petrologic studies, in the search for economic PGE concentrations of sulfide-poor PGE mineralization in other volcanic and possibly plutonic rocks as well. Specifically, the results suggest that podiform magnetic anomalies within titaniferous magnetite-rich pyroxenites and gabbroic rocks may have potential use in the exploration for economic sulfide-poor PGE mineralization. Although the paleomagnetic methods used in this study are probably not of direct use in exploration, they were able to distinguish the different lithologies in the BCF. This substantiates the results of the susceptibility measurements in characterizing PGE mineralized, titaniferous magnetite-bearing rocks. The results of the paleomagnetic study also show that the NRM of the Ghost Range intrusive complex is not primary and therefore the Archean apparent polar wander path as it is currently defined is incorrect.</p> / Thesis / Bachelor of Science (BSc)

The investigation of isomeric states in neutron-rich nuclei provides useful insights into the underlying nuclear configurations, and understanding their occurrence along an isotopic chain can inform about shell evolution. Recent studies on neutron-rich Si isotopes near the magic number N = 20 and approaching N = 28 have revealed the presence of low-lying states with intruder configurations, resulting from multiple-particle, multiple-hole excitations across closed shell gaps. The characterization of these states involves measuring their half-lives and transition probabilities. In this study, a new low-energy (7/2−1) isomer at 68 keV in 37Si was accessed via beta decay and characterized. To achieve this, radioactive 37Al and 38Al ions were produced through the projectile fragmentation reaction of a 48Ca beam and implanted into a CeBr3 detector, leading to the population of states in 37Si. The 68-keV isomer was directly populated in the beta-delayed one neutron emission decay of implanted 38Al ions. Ancillary detector arrays comprising HPGe and LaBr3(Ce) detectors were employed for the detection of beta-delayed gamma rays. The choice of detectors was driven by their excellent energy and timing resolutions, respectively. The beta-gamma timing method was utilized to measure the half-life of the new isomeric state in 37Si. This dissertation also discusses other timing techniques employed to search for and characterize isomeric states following beta decay of implanted ions. Notably, the half-life of the newly observed (7/2−1) isomeric state in 37Si was measured to be 9.1(7) ns. The half-life of the previously observed closely-lying (3/2−1) state at 156 keV was determined to be 3.20(4) ns, consistent with previously reported values. Reduced ground-state transition probabilities associated with the gamma-ray decay from these excited states were in agreement with results obtained from shell model calculations. In addition to the investigation of isomeric states in 37Si, isomeric 0+ states in 34Si and 32Mg nuclei belonging to the N = 20 “island of inversion” were characterized and searched for, respectively. The isomeric 0+ state in 34Si was populated following the beta decay of implanted 34Mg ions and its 34Al daughter nucleus. Similarly, the 0+ state in 32Mg was searched for via the beta-delayed one neutron emission decay of implanted 33Na ions.

Recent studies suggest that factor (F) XII plays a key role in thrombus stabilization and growth but is dispensable for hemostasis. We have previously shown that histidine-rich glycoprotein (HRG), a protein present in platelets and plasma, binds FXIIa and inhibits FXII autoactivation and FXIIa-mediated activation of FXI, thereby downregulating thrombosis. HRG binds various ligands, including FXIIa, fibrin(ogen), nucleic acids and polyphosphate (polyP). Studies have shown that polyP, released from activated platelets, and artificial surfaces like catheters, can promote FXII activation. This suggests that HRG can downregulate the activation of the contact system. This thesis aims to determine the potential mechanisms by which HRG modulates platelet function and thrombosis induced by polyP or catheters. We show that HRG binds polyP with high affinity and inhibits the procoagulant, prothrombotic and cardiotoxic effects of polyP via at least two mechanisms. First, HRG binds polyP and neutralizes its procoagulant activities and cytotoxic effects. Second, HRG binds FXIIa and attenuates its capacity to promote autoactivation and activate FXI. Also, we identify that HRG serves as a molecular brake for the contact system by attenuating the procoagulant activity of FXIIa regardless of whether FXII activation is triggered systemically with polyP or occurs locally on the surface of catheters. Our studies have identified HRG as a novel ligand for platelet receptor GPIbα on resting platelets, and upon activation, it competes with fibrinogen for binding to GPIIb/IIIa integrin, thereby inhibiting platelet aggregation. These findings suggest that HRG may modulate coagulation as well as platelet function. Therefore, supplementation with HRG or HRG analogs may serve as a potential therapeutic option to attenuate polyP or catheter-induced thrombosis without perturbing hemostasis. / Dissertation / Doctor of Philosophy (PhD)

At ~662 gigatonnes of carbon (GtC), marine dissolved organic matter (DOM) is the largest reduced pool of actively cycling carbon and nitrogen in the oceans1. Operationally defined as smaller than 0.1µm in size, this carbon reservoir comprises all non-living organic matter smaller than a bacterial cell and comprises organic colloids and molecules spanning as a continuum of sizes ranging from marine viruses and large macromolecules (e.g. DNA, enzymes) to small organic molecules (e.g. polymers and monomers)2. With deep apparent 14C-ages ranging between 4900-6400 ybp 3,4, marine DOM is anomalously old given timescales of global ocean ventilation (1000-1500 years). The great age of DOM has remained one of the most elusive lines of scientific inquiry in Chemical Oceanography for decades. The size and molecular composition of DOM has been shown to be a key variable in determining its biological reactivity (e.g. cycling rate) and long-term persistence in the deep ocean5,6. Despite the importance of DOM in the marine carbon and nitrogen cycles, we lack a detailed understanding of the molecular composition of DOM. Due to the high concentration of salts in seawater relative to DOM, it is difficult to analyze the molecular composition of seawater with conventional chemical- or size- fractionation methods without introducing bias (i.e. isolating only hydrophobic and/or high molecular weight DOM). In fact, it is commonly reported that >80% of DOM remains uncharacterized at the molecular level (e.g. not readily identifiable as an individual known biomolecule)5. Nuclear magnetic resonance (NMR) spectroscopy has been used as a tool for several decades to describe the composition of marine DOM isolates7. For example, 13C-NMR of major high molecular weight DOM functional groups at the molecular-level demonstrated that DOM is largely made up of reactive polysaccharides with low aromaticity compared to terrestrial DOM8. To date, all marine DOM NMR measurements have been made on size-fractionated DOM or chemically-fractionated (e.g. solid phase extracted) DOM isolates. In this thesis, I report the first Proton (1H) NMR composition of total seawater DOM from seawater samples collected from 10 stations in Baffin Bay aboard the CCGS Amundsen (2019). Samples were measured using 1H-NMR at uOttawa following a novel water suppression method established by Lam and Simpson9. The use of this method has allowed for the first molecular composition assessment of total seawater DOM to be measured (e.g. without any chemical or size fractionation). I report the % relative abundance of individual biomarkers and determine molar concentrations of two compound classes of interest. These results are shown in Ocean Data View section plots, and are listed within appendix tables, to provide a comprehensive depiction of the changing concentrations of dissolved organic carbon (DOC), total carbohydrates (TCHO), and carboxyl-rich alicyclic molecules (CRAM). In this thesis, I explore changes in the abundance of these unique DOM compound classes and discuss how the composition of DOM directly determines its bioavailability and thus cycling in Baffin Bay 5. The core objective of my thesis was to measure DOM concentrations for TCHO and CRAM, as well as to calculate the production and removal of these key DOM compounds in Baffin Bay due to either physical and/or biological processes. We found that the concentration of both TCHO and CRAM decreased with depth throughout Baffin Bay. This is consistent with previous work suggesting the rapid cycling of carbohydrates, however it contradicts the current paradigm of CRAM cycling. Our results indicate between 21-43% of CRAM produced in the surface is subsequently removed at depth. Rapid cycling of a surface CRAM population suggests that not all CRAM can be considered recalcitrant DOM We live in a time of unprecedented global change. The Arctic Ocean is warming at a rate at least four times faster than the global average10. The impact of a rapidly warming, freshening and increasingly acidified Arctic Ocean on the biogeochemistry of DOM remains unknown. It is imperative that more DOM research be conducted as early as possible in order to better understand these impacts and inform future research directions. The distribution and cycling of CRAM in Baffin Bay provide novel and fundamental knowledge of DOM cycling in a key Arctic region, but could also potentially occur throughout the global ocean. Such data will no doubt be of use in informing future iterations of Earth System Climate models seeking to forecast how the marine carbon cycle will respond to global change.