Scholar Commons - Institutional Repository of the University of South Carolina
Not a member yet
34867 research outputs found
Sort by
Between Healing and Harm: An Analysis of Residential Treatment Centers for Youth
Full text linkResidential treatment for children and adolescents has evolved greatly. Today, a therapeutic individualized treatment approach is standard. Despite past and present improvements in care, this paper examines the persistent challenges faced by residential treatment centers (RTCs) serving youth with mental health issues. This paper argues that residential treatment, while sometimes necessary, has the potential to cause more harm than good due to systemic issues like overdiagnosis, inappropriate medication, and the troubled teen industry’s (TTI) profit-focused approach
Tsenacommacah
No full textThis poem is about the area that was once called Tsenacommacah, the Powhatan people, and the horrors of the centuries that followed 1607
Glimpses of Myself in the Underground
No full textThrough poetry, a older teenager reflects on a place that has made her who she is. She sees ghosts of her past self in every wall of this place, and it makes her feel oddly sentimental. She realizes how far she has come and how much has changed, and begins to understand how ever fleeting any given moment in time is
The Dischargeability of Student Loans Through Bankruptcy: Does the Undue Hardship Standard Adversely Burden Consumer Debtors?
No full text
No Child Left Behind Bars: The Educational Attainment of Children with Incarcerated Parents
No full textAssessing Bioactivity of Green Leaf Volatiles in Various Plant Systems
No full textGreen leaf volatiles (GLVs) are chemical compounds essential in plant communication. Plants emit these volatiles due to abiotic and biotic stressors and, when perceived, these GLVs activate defense responses. Though GLVs play an important role in plant survival, it is still unknown where they localize, how they are perceived, and what structural factors influence perception. To elucidate these structural determinants, we tested the bioactivity of (Z)-3-fatty alcohols with four to nine carbons. Through root growth inhibition assays with tomato (Solanum peruvianum) seedlings, we found that (Z)-3-octenol and (Z)-3-nonenol were more bioactive than (Z)-3 hexenol and (Z)-3-heptenol. Thus, higher bioactivity was seen among the longer chain lengths. The exception was (Z)-3-butenol, which had a higher bioactivity than (Z)-3-hexenol. To determine intracellular signaling events upon GLV perception, we tested for the phosphorylation of MAP kinases (MAPKs) in Arabidopsis (Arabidopsis thaliana). We found that both HOL and HAC induce MAPK3/6 phosphorylation in different application systems with seedlings. We also wanted to determine if adding an azide mini tag onto the GLV (Z)-3-hexenyl acetate would impact the bioactivity. In Solanum peruvianum suspension-cultured cells, we found that both the (Z)-3-hexenyl acetate and the mini-tagged derivative induce medium acidification. However, in Arabidopsis seedlings, we found that only (Z)-3-hexenyl acetate induced a MAP kinase phosphorylation response. We also tested the bioactivity of azide-tagged GLV derivatives with six to nine carbons to explore how the carbon chain length influences bioactivity. We found that the longer the carbon chain, the lower the medium acidification response. Thus, there is a reverse relationship between bioactivity and chain length for mini-tagged (Z)-3-hexenyl acetate derivatives
Of Grave Importance: A Collaborative Approach to Archaeological Research on the Cooper River Cemetery, Daufuskie Island, South Carolina
No full textLand dispossession and displacement is a challenge being faced by the Gullah community in Lowcountry South Carolina. Land dispossession due to a phenomenon known as coastal capitalism, or, the increased interest in coastal property, on Daufuskie Island served as an impetus for Miss Sallie Ann Robinson, a Gullah, Daufuskie native, to reach out to academic communities to assist in the preservation of spaces including cemeteries. This led to the ongoing collaborative archaeology project on the Cooper River Cemetery to record information about burials and investigate the presence of those without any headstone markers. The Cooper River Cemetery is one of five Gullah Cemeteries on the island and is part of a larger research initiative set up by the Daufuskie Island Gullah Heritage Society. Using Ground Penetrating Radar, GIS and archival maps, the data has been analyzed through the lens of necrogeography, with the Cooper River Cemetery as the deathscape of focus. This research highlights the importance of collaboration within the field of archaeology, and considers the ways that archaeologists can contribute to, and assist in answering questions posed by communities
New Deep Learning Approaches to Classical Statistical Problems
Full text linkThe field of deep learning (DL) has received considerable attention in recent years. Thanks to rapid growth in computational power, the ability to collect massive datasets, and improvements in software and algorithms, DL is now routinely applied to areas as diverse as computer vision, natural language processing, and bioinformatics. At the same time, DL is only starting to be explored in the context of classical statistical inference problems such as bootstrapping, quantile regression, and mixture modeling. In this dissertation, we develop new DL methodology for three classical statistical problems: 1) weighted M-estimation, 2) joint quantile regression, and 3) mixing density estimation.
In Chapter 2, we introduce a deep learning generative framework for efficient weighted M-estimation. To overcome computational bottlenecks of various data perturbation procedures such as the bootstrap and cross-validation, we propose the Generative Multi-purpose Sampler (GMS), which directly constructs a generator function to produce solutions of weighted M-estimators from a set of given weights and tuning parameters. The GMS is implemented by a single optimization procedure without having to repeatedly evaluate the minimizers of weighted losses, and is thus capable of significantly reducing the computational time. We demonstrate that the GMS framework enables the implementation of various statistical procedures that would be infeasible in a conventional framework, such as iterated bootstrap procedures and cross-validation for penalized likelihood. To construct a computationally efficient generator function, we also propose a novel form of neural network called the weight multiplicative multilayer perceptron to achieve fast convergence.
In Chapter 3, we introduce a deep learning generative model for joint quantile estimation called Penalized Generative Quantile Regression (PGQR). Our approach simultaneously generates samples from many random quantile levels, allowing us to infer the conditional distribution of a response variable given a set of covariates. Our method employs a novel variability penalty to avoid the problem of vanishing variability, or memorization, in deep generative models. Further, we introduce a new family of partial monotonic neural networks (PMNN) to circumvent the problem of crossing quantile curves. A major benefit of PGQR is that it can be fit using a single optimization, thus bypassing the need to repeatedly train the model at multiple quantile levels or use computationally expensive cross-validation to tune the penalty parameter.
In Chapter 4, we propose a deep generative process for mixing density estimation in latent variable models called Generative Bootstrapping for Nonparametric Maximum Likelihood Estimation (GB-NPMLE). GB-NPMLE rapidly produces NPMLE bootstrap estimates for estimating an unknown continuous mixture distribution. Traditional bootstrapping requires repeated evaluations on resampled data and is not scalable. On the other hand, GB-NPMLE requires only a single evaluation of a novel two-stage optimization algorithm. Our procedure accurately estimates continuous mixing densities with little computational cost even when there are a hundred thousand observations.
In Chapter 5, we introduce neural-g, a new neural network-based estimator for mixing density estimation. Neural-g uses a softmax output layer to ensure that the estimated prior is a valid probability density. Under default hyperparameters, we show that neural-g is very flexible and capable of capturing many unknown densities, including those with flat regions, heavy tails, and/or discontinuities. We provide theoretical justification for neural-g by establishing a new universal approximation theorem regarding the capability of neural networks with softmax output layers to learn arbitrary probability mass functions. To accelerate convergence of our numerical implementation, we utilize a weighted average gradient descent approach to update the network parameters. Finally, we extend neural-g to multivariate prior density estimation