21793 research outputs found
Sort by
Descent: A Novella
No full textThe novella follows Yanet, a 23-year-old Film Production student and aspiring filmmaker living in Toronto, as she struggles with ebbing friendships and artistic failures after being put on academic probation at her university a year after the death of her father. The story is told in first-person narration to reflect Yanet’s stream-of-consciousness about her experience and the various relationship dynamics between her close friends, her Colombian immigrant parents, and her siblings. The title gestures to the narrative tailspin following Yanet’s academic probation and to one’s familial and cultural origins.
The coming-of-age story explores a number of overlapping themes: grief, failure, artistic aspirations, and the formation of a cultural identity for children of immigrants in Toronto. Her friends, Paola and Toni, are students at the same university, and they are in their final term before graduating. A fourth friend, Michelle, has already graduated, and Yanet attends the release of Michelle’s first film.
The narrative follows the protagonist’s endeavor to complete her capstone project, a documentary of female friendship, among a burgeoning arts community in Toronto, and as she attends the anniversary mass of her father’s death. Divided into three parts, the novella’s structure symbolizes how Yanet partitions her life: her social life downtown and her family life are kept separate, with her friend, Paola, being the only link between the two.
Something of an anti-Künstlerroman, Yanet struggles to fulfill her artistic potential and fails in many ways
Ordering Problems in Natural Language Generation
Get PDFLanguage is the main mechanism used in human communication. For decades, researchers
have tried to build computers capable of communicating with humans in natural languages. Such research endeavors include building systems to understand human language (a.k.a Natural Language Understanding or NLU) and generate a response in human language (a.k.a Natural Language Generation or NLG). The goal of NLG is to deliver information by producing natural texts in human languages. Text generation typically consists of several steps in a pipeline, from determining the content to communicate to producing the actual words, and hence requires different techniques at each step. In this thesis, we focus on ordering problems in NLG and address two tasks that require ordering; namely, sentence ordering and surface realization.
In sentence ordering, models need to capture the relations between sentences and then based on these relations, find the most coherent order of sentences. Our proposed approach is based on pointer networks where at each step, the model chooses the sentence that should appear next in the text. We show that using a conditional sentence representation that captures the sentence meaning based on its position and previously selected sentences in the text can improve the state-of-the-art on standard datasets.
For surface realization, we show that a pointer network is insufficient to improve state-of-the-art performance. Therefore, we propose the use of language models for surface realization by mapping the surface realization task (a graph-to-text) to a text-to-text problem. Our experiments show that pre-trained language models can easily learn the task of surface realization and achieve competitive performances on standard datasets. To further improve the performance of surface realization, we then propose to pre-train a language model on synthetic data and then fine-tune it on manually labeled data in order to increase the number of training data for surface realization. Our experiments indicate that this approach improves the state-of-the-art by more than 10% BLEU score on standard datasets
reduction responsive superparamagnetic iron oxide nanoparticle clusters for t2-t1 magnetic resonance imaging
Get PDFThe main objective of this M.Sc. research is to synthesize a novel stimuli-responsive (SR) poly(acrylic acid) (PAA)-stabilized ultra-small superparamagnetic iron oxide nanoparticle (USNP) clusters as a switchable T2-T1 contrast agent for magnetic resonance imaging (MRI). Oleic acid (OA)-stabilized USNPs were synthesized in an organic solvent and their biphasic ligand exchange with an aqueous PAA solution yielded an aqueous dispersion of PAA stabilized USNP colloids (called PAA-USNPs) with the hydrodynamic diameter of ≤ 20 nm. The former colloids reacted with cystamine (Cys) to form disulfide-labeled PAA-USNP clusters with a diameter of >100 nm through the formation of amide linkages between terminal amino groups of Cys and carboxylic acid groups on the PAA-USNP. The resultant clusters reverted to individual PAA-USNPs with a hydrodynamic diameter close to that (≤ 20 nm) of PAA-USNPs upon the cleavage of disulfide bonds in the presence of a reducing agent such as dithiothreitol or glutathione. These SR PAA-USNP clusters have promising functions as an effective T2-T1 contrast agent. It can be anticipated that these PAA-USNPs and their clusters could find potential biomedical applications for the dual monitoring of T1 and T2 contrast using MRI of the bladder
QAnon: A Survey of the Evolution of the Movement from Conspiracy Theory to New Religious Movement
Get PDFDisinformation and conspiracy theories have quickly developed into threats against democratic institutions, violent extremism, threats against elected leaders, and attacks on vital infrastructure. None has had a greater influence on the violent extremist community than QAnon, which emerged from the Pizzagate hoax and has developed over the past five years into a significant ideology motivated violent extremist movement. The COVID-19 epidemic has strengthened xenophobic and anti-authority narratives, many of which may have an adverse effect on national security, democratic institutions, and public health. This has contributed to its quick emergence as a significant threat actor. By spreading fake information about government actions and the virus itself online and distorting the truth to support its conspiracy theory and worldview, QAnon took advantage of the pandemic. In an effort to excuse and justify killing, some violent extremists in QAnon have accepted conspiracy theories concerning the pandemic. These tales have helped to erode confidence in both scientific knowledge and the honesty of the government. While some conspiracy theory rhetoric is a valid exercise in the right to free speech, internet discourse that is becoming more aggressive and demands for the detention and killing of particular people raises serious concerns.
This dissertation will represent an initial examination of some of the facets of the QAnon movement by examining the complex history of the QAnon as it has fluctuated and evolved, not only due to the likelihood of multiple users behind the “Q” account, but also as the sociopolitical landscape has changed since the creation of the movement. This dissertation will (1) frame QAnon as a lived religion and demonstrate that it has gone through three stages of existence: proto-QAnon, canonical-QAnon and apocryphal-QAnon. (2) It will then argue that QAnon overtimed evolved into something more than a conspiracy theory but will argue in a comparative analysis that QAnon is more akin to a new religious movement, in particular a hyper-real religion. (3) Next it will examine the role of gender and women in the QAnon movement. (4) This will be followed by an examination of how the QAnon conspiracy theories have legitimized, coordinated an targeted gender based violence. (5) Will provide evidence of the nexus of QAnon and ideologically motivated violent extremism and criminality. (6) Finally, it will examine the evolution of QAnon after the January 6th insurrection, the lost of the election by Donald Trump and the disappearance of “Q”
Investigation of the Influences on Surface Resistivity used for Quality Control of Concrete
Get PDFThe destructive techniques that can be used to estimate the concrete quality are expensive and time consuming and are often accompanied by following reinforcing and repair. Therefore, non-destructive testing (NDT) methods have drawn the attention of some researchers. There is an increasing interest in electrical resistivity as non-destructive tests for quality control of concrete structures in the last decades. Electrical resistivity is a technique that is rapid and inexpensive. This project studied how the electrical resistivity of concrete and mortar cylinders immersed in different solutions changed with time in order to investigate various factors. In the two rounds tests of concrete and mortar specimens, it was found that the electrical resistance is highly influenced by the moisture content of the specimens, the immersion solutions, the water-cement ratio of the concrete as well as the addition of supplementary cementitious materials (SCMs). In addition, it was observed in the experiments that the electrical resistivity of the concrete specimens decreased rapidly within 24 hours after immersion in the solution (the specimens were dried in advance for 30 days), and the rate of decrease gradually slowed down over the next 27 days. It is also showed that the electrical resistivity of mortar specimens was significantly lower than the concrete with a similar mixture design. However, for the portion of the research that investigated the solution penetration with time by using internal sensors, the tests results are not adequate data to draw a conclusion
Essays in Corporate Finance, Shareholder Litigation, and Politics
Get PDFWhen firms seek to curry the favor of politicians, it inevitably leads to political corruption. Political spending totaled US$14.4 Billion in the 2020 US election cycle—and this total does not include dark money donations. Firms naturally never donate to politicians without wanting a return on their investment, so clearly political corruption is a multi-billion-dollar problem in the United States. Recently, a strand of literature examines political corruption in the US from a corporate finance perspective. Another recent strand of finance literature concerns the effects of political ideology on the outcomes of US securities-related shareholder litigation. This thesis aims to first combine and expand upon these two emerging strands of literature by analyzing the relationships of a comprehensive variety of US political and judicial variables with the outcomes of securities fraud and related shareholder litigation. We then extend our framework to a refined exploration of corporate governance as it relates to shareholder litigation.
In the first essay, we study the relationship between a number of political and judicial variables in the United States with the outcomes of litigation for firms that have been sued by their shareholders. Consistent with our hypothesis, we find that a crucial factor in shareholder litigation dismissal has been the passage of the Citizens United v. FEC Supreme Court campaign finance ruling of 2010. Furthermore, we find evidence that political campaign contributions afford firms the requisite connections that will benefit them in current or future lawsuits. Also, we quantify the impact of the size and timing of the political campaign contributions. In addition, we confirm hypotheses that the fate of shareholder class action litigation against these firms is also affected by the political ideologies of some of the trusted authorities who write, administer, and interpret the laws pertinent to firms facing such litigation. These authorities are federal politicians and judges, who are ideally independent arbiters—but the great powers they are given appear to create agency and bias issues, respectively.
In the second essay, we use the knowledge and framework attained from our conclusions from the first essay to examine various corporate governance variables with respect to their role in shareholder litigation outcomes in this new light—variables which can be categorized as board, executive, and firm ownership characteristics. We confirm hypotheses generally based on the principle that variables reflecting better corporate governance will tend to be associated with a higher lawsuit dismissal likelihood. This likelihood tends to increase with a firm’s board of directors who are older, more independent, less busy, and have a larger network size. Furthermore, the likelihood of litigation dismissal increases with greater analyst coverage of the firm, with a firm’s CEO who is older than the board of directors, with greater institutional ownership, and with a larger number of blockholders owning stakes in the firm. As well as finding results consistent with such hypotheses for our corporate governance variables, we also find some novel, unexpected interactions between political variables and corporate governance variables
At What Point(s) in the Research Lifecycle Can an Art Library Facilitate Visual Art Research?
No full textThis presentation looks at the question “At what point(s) in the research lifecycle can an art library facilitate visual art research?” in relation to an ongoing research-creation project that I began during a spring/summer residency at Artexte in 2018, and that will culminate in a publication to be co-published by this non-profit Canadian arts organization and myself in 2023.
Entitled Who Was Who Was Who in Contemporary Canadian Art, the residency and publication explore and document Canadian artists from the 1960s onwards who use pseudonyms, personae, alter egos and other kinds of alternate identities in their art practice. This bilingual (English/French) publication takes the form of a print and openly accessible artists’ biographical dictionary with distinct but related entries for the artists and their alternate identities.
Through the use of concrete examples in my presentation, I will argue that an art library like Artexte’s can facilitate visual art research throughout the entire research lifecycle in numerous and invaluable ways
Development and Application of Children's Sex- and Age-Specific Fat-Mass and Muscle-Mass Reference Curves using the LMS Methodology
Get PDFBody mass index cannot distinguish between fat-mass and muscle-mass, which may result in obesity misclassification. A dual-energy x-ray absorptiometry (DXA)-derived phenotype classification based on fat-mass and muscle-mass has been proposed for adults (>18 yo). We extend this research by developing children’s fat-mass and muscle-mass reference curves and determining their utility in identifying cardiometabolic risk. Children’s (≤17 yo) DXA data in NHANES, a US national health survey (n=6,120) were used to generate sex- and age-specific deciles of appendicular skeletal muscle index and fat mass index (kg/m2) with the Lambda Mu Sigma (LMS) method. The final curves were selected through goodness of fit (AIC, Q-tests, detrended Q-Q plot). Four phenotypes (high [H] or low [L], adiposity [A] and muscle mass [M]: HA-HM, HA-LM, LA-HM, LA-LM) were identified using the literature’s guidelines above/below the median compared to same-sex and same-age peers. The curves and their corresponding phenotypes were applied to QUALITY data, a longitudinal cohort (n=630, 8-10 yo in 2005) to assess whether the phenotypes correctly identified cardiometabolic risk using multiple linear regression at baseline, follow-up one (2008-2010), or follow-up two (2015-2017). Models were adjusted for age, sex, and Tanner’s stage. Chained equation was used to impute missing values in QUALITY. Compared to LA-HM, LA-LM was associated with lower glucose at baseline; HA-HM was associated with lower HDL-c and higher LDL-c, triglycerides, and HOMA-IR; HA-LM was associated with elevated triglycerides and HOMA-IR at all timepoints (all p<0.05). These phenotypes allowed for discrimination of cross-sectional cardiometabolic risks, but further longitudinal exploration is recommended
Unsupervised Learning with Feature Selection Based on Multivariate McDonald’s Beta Mixture Model for Medical Data Analysis
Get PDFThis thesis proposes innovative clustering approaches using finite and infinite mixture models to analyze medical data and human activity recognition.
These models leverage the flexibility of a novel distribution, the multivariate McDonald’s Beta distribution, offering superior capability to model data of varying shapes. We introduce a finite McDonald’s Beta Mixture Model (McDBMM), demonstrating its superior performance in handling bounded and asymmetric data distributions compared to traditional Gaussian mixture models.
Further, we employ deterministic learning methods such as maximum likelihood via the expectation maximization approach and also a Bayesian framework, in which we integrate feature selection. This integration enhances the efficiency and accuracy of our models, offering a compelling solution for real-world applications where manual annotation of large data volumes is not feasible.
To address the prevalent challenge in clustering regarding the determination of mixture components number, we extend our finite mixture model to an infinite model. By adopting a nonparametric Bayesian technique, we can effectively capture the underlying data distribution with an unknown number of mixture components.
Across all stages, our models are evaluated on various medical applications, consistently demonstrating superior performance over traditional alternatives.
The results of this research underline the potential of the McDonald’s Beta distribution and the proposed mixture models in transforming medical data into actionable knowledge, aiding clinicians in making more precise decisions and improving health care industry
Orchard Apple Tree Health Assessment using UAV Imagery-Based Computer Vision System
Get PDFAccurate and efficient orchard tree inventories play a crucial role in obtaining up-to-date information for effective treatments and crop insurance purposes. Surveying orchard trees, including counting, locating, and assessing their health status, is vital for predicting production volumes and
facilitating orchard management. However, traditional manual inventories are labor-intensive, expensive, and prone to errors. Motivated by the recent advances in UAV imagery and computer
vision methods, we propose a new framework for individual tree detection and health assessment.
The proposed approach follows a two-stage process. First, we build a tree detection model based
on a hard negative mining strategy using RGB UAV images. In the second stage, we address the
health classification problem using two methods. We present a classical machine learning approach
by exploring the use of multi-band imagery-derived vegetation indices. We also propose a new convolutional autoencoder-based architecture mainly designed to extract the relevant features for tree
health classification.
The performed experiments demonstrate the robustness of the proposed framework for orchard
tree health assessment from UAV images. In particular, our framework achieves an F1-score of 86.24% for tree detection and an overall accuracy of 98.06% for tree health assessment. Moreover, our work could be generalized for a wide range of UAV applications involving a detection/classification process