Electronic Thesis or Dissertation

Neural network models with many tunable parameters can be trained to approximate functions that transform a source distribution, or dataset, into a target distribution of interest. In contrast to low-parameter models with simple governing equations, the dynamics of transformations learned in deep neural network models are abstract and the correspondence of dynamical structure to predictive function is opaque. Despite their “black box” nature, neural networks converge to functions that implement complex tasks in computer vision, Natural Language Processing (NLP), and the sciences when trained on large quantities of data. Where traditional machine learning approaches rely on clean datasets with appropriate features, sample densities, and label distributions to mitigate unwanted bias, modern Transformer neural networks with self-attention mechanisms use Self-Supervised Learning (SSL) to pretrain on large, unlabeled datasets scraped from the internet without concern for data quality. SSL tasks have been shown to learn functions that match or outperform their supervised learning counterparts in many fields, even without task-specific finetuning. The recent paradigm shift to pretraining large models with massive amounts of unlabeled data has given credibility to the hypothesis that SSL pretraining can produce functions that implement generally intelligent computations.

Mythological stories are more than stories. They are important because they give context to our world, our literature, and our own beliefs and carry forward cultural values and tales over generations. So many efforts have been made to pass mythological stories from generation to generation through various forms including traditional storytelling, songs, and films. This thesis explores if learning mythology can be more immediate and fun by instructing it through an immersive, interactive game known as an Escape Room. Riddles of the Gods is an Egyptian theme-based Escape Room that uses the medium of Virtual Reality. It creates an interactive, educative, and entertaining experience for participants. It also provides an opportunity to contribute towards keeping traditional and cultural values alive, in a more expressive form using the latest technologies such as Virtual Reality which are relatable to the current and future generations.

Serine/Arginine splicing factor 1 (SRSF1) is an RNA-binding protein (RBP) with multiple functions in RNA biogenesis. SRSF1 plays a prominent role in oncogenesis, immune function, and response to several physiological stimuli. To date, the role of SRSF1 as a regulator of mRNA splicing has been largely considered the main mechanism driving its biological functions and its role in disease. We have now characterized SRSF1’s role in Human Immunodeficiency Virus Type I (HIV-1) transcription. SRSF1 interacts with the 7SK small nuclear ribonucleoprotein (snRNP) to mobilize and activate the positive transcription-elongation factor (P-TEFb), which is then positioned on the HIV-1 promoter to increase the processivity of RNA polymerase II (RNAPolII) and promote the release of the negative regulators of transcription DSIF/NELF. Next, we defined the role of SRSF1 in the transcription of cellular genes utilizing an RNA sequencing (RNASeq) time course approach was used to detect changes in the transcriptome in response to SRSF1 overexpression. RNASeq data analysis revealed a subset of genes that were upregulated in response to SRSF1 overexpression. Nuclear run-on and qPCR assays experimentally validated 28 of these genes.

This thesis aims to contribute to contemporary feminist theory through the integration of several interdisciplinary texts from the last century, all of which challenge an existing, male-oriented norm of woman as ‘lesser’ in a particular field of study. The historical position of woman as ‘other’ in a negative light is a postulate that contemporary feminist studies may take too much for granted. The supposed lack of prominence of women in scripture, such as what Phyllis Trible gestures to, for example, is not erasure at all, but women present as archetypes, a mode of representation later dispersed in literature and film. The textual ‘absence’ of the feminine which has been previously understood as erasure may in fact be a clandestine interpretative tool which must be sought for, or, within a textual framework, explicated. Instead of accepting woman as a minimized ‘other’ to be merely a given in biblical and other texts, her peripheral role must be teased out in order to be fully appreciated. The critical most important to this claim include Carol Gilligan’s In a Different Voice: Psychological Theory and Women’s Development and film theorist Molly Haskell’s From Reverence to Rape: The Treatment of Women in the Movies, the latter of which lends this thesis its title. Lastly, I will be using erasure as an interpretative method as applied to a series of case studies: to analyze the female figures in Hamlet using Carol Gilligan’s psychological development framework; to consider Haskell’s rigorous critique of American cinema alongside Woman in the Dunes, a 1964 film based on a fabulist novel, which uses erasure as its modus operandi; and to apply Phyllis Trible’s hermeneutic interpretive method to Lot’s wife. The interdisciplinary design of this thesis allows for the inclusion of scholars from a variety of inherently ethical disciplines to showcase how societal perceptions of women have informed women’s ethical decision-making and identity.

In radiotherapy, radiobiological indices tumor control probability (TCP), normal tissue complication probability (NTCP), and equivalent uniform dose (EUD) are computed by analytical models. These models are rarely employed to rank and optimize treatment plans even though radiobiological indices weights more compared to dosimetric indices to reflect treatment goal. The objective of this study is to predict TCP, NTCP and EUDs for lung cancer radiotherapy treatment plans using an artificial neural network (ANN). A total of 100 lung cancer patients’ treatment plans were selected for this study. Normal tissue complication probability (NTCP) of organs at risk (OARs) i.e., esophagus, spinal cord, heart and contralateral lung and tumor control probability (TCP) of treatment target volume (i.e., tumor) were calculated by the equivalent uniform dose (EUD) model. TCP/NTCP pairing with corresponding EUD are used individually as outputs for the neural network. The inputs for ANN are planning target volume (PTV), treatment modality, tumor location, prescribed dose, number of fractions, mean dose to PTV, gender, age, and mean doses to the OARs. The ANN is based on Levenberg-Marquardt algorithm with one hidden layer having 13 inputs and 2 outputs. 70% of the data was used for training, 15% for validation and 15% for testing the ANN. Our ANN model predicted TCP and EUD with correlation coefficient of 0.99 for training, 0.96 for validation, and 0.94 for testing. In NTCP and EUD prediction, averages of correlation coefficients are 0.94 for training, 0.89 for validation and 0.84 for testing. The maximum mean squared error (MSE) for the ANN is 0.025 in predicting the NTCP and EUD of heart. Our results show that an ANN model can be used with high discriminatory power to predict the radiobiological indices for lung cancer treatment plans.

Kersten et al. (2021) revealed that participants remembered negatively valenced actions better than neutral actions, but did no better at binding negative actions with the people who performed them compared to neutral actions. We were interested in testing whether emotion only enhances memory for individual features of an event, or whether emotion can also enhance binding of certain combinations of features. In particular, we tested the effect of emotionally charged objects on the ability to remember those objects and the actions associated with them. Participants saw a series of brief videos each involving an actor performing one of two different actions on one of two objects within a specific object category (e.g., guns or piñatas), some objects neutral in valence, some positive, and some negative. Participants were later tested on their ability to distinguish old events from novel conjunctions of particular objects with the actions that had been previously performed with the other members of the same object categories. Although only marginally significant, participants appeared more able to bind objects with their associated actions when those objects held a negative charge compared to neutral objects. Additionally, participants were more sensitive to changes in actions when those actions were associated with negative objects compared to neutral or positive objects. However, false memory increased when new negative objects were presented compared to novel presentation of neutral or positive objects.

Nuisance odors from landfills have more impact than just being an annoyance to nearby residents. With an ever-increasing population, a larger number of communities are located in closer proximity to landfills than ever before. This has brought along with it, more regular conflicts with landfill authorities surrounding the issue of odors, resulting in complaints, lawsuits, fines, and even re-siting operations. The absence of an objective method of quantifying nuisance odors makes the task of creating regulations and setting standards even more complicated. The current research focuses on a method to objectively quantify landfill odors. The human odorant binding protein 2A (hOBPIIa) can be produced using published recombinant gene technology and can be used as a biosensor to quantify odorants through spectrofluorometric measurements. The current work is a continuation of the previous work by Rahman (2020). In this work, the spent biosensor after it reacts with an odorant is shown to be regenerated by applying additional fluorophore following La Chateliers’ principle, so that the same batch of protein can be used to run multiple experiments with odorants. An important part of the work miniaturized the earlier version of the experimental setup and incorporates a much more efficient flow-through system. This setup is capable of collecting real-time readings, increasing the overall accuracy and shortening the duration of each set of the experiment. The current work also explores the response of the biosensor with an expanded group of pure odorants, including hydrogen sulfide, ammonia, toluene, formaldehyde, tert-butyl mercaptan, and methyl mercaptan as well as their mixtures, thus expanding the list of odorants tested under this principle. The results show that the protein shows a concentration-dependent response differing on the hydrophobicity of the target compound.

Exosomes have gained recognition in cancer diagnostics and therapeutics. Most exosome isolation methods are time-consuming, costly and require bulky equipment, rendering them unsuitable for point-of-care (POC) settings. Microfluidics can be the key to solving these challenges. Here, we employ the development of a double filtration microfluidic device that can rapidly isolate exosomes in POC settings. The device can efficiently isolate exosomes from just 100 uL of plasma within 50 minutes. The device was compared against Polyethylene glycol (PEG) based precipitation, and findings show that both methods yield comparable exosome sizes and purity, but the device can detect exosomal miRNA earlier than PEG. Finally, a comparative analysis of membrane filters with exosomes collected from pore sizes 15 nm and 30 nm showed a similarity in exosome size and miRNA expressions, with significantly increased sample purity. These findings suggest that this device has potential in POC settings.

In this effort, we present progress toward demonstrating a Decoy-State Quantum Key Distribution (QKD) source based on a polarization-modulator and a wavelength-stable attenuated pulsed laser. A three-state QKD protocol is achieved by preparing particular quantum polarization states. The polarization-modulator-based QKD source improves security by removing several sources of side-channel attacks that exist when multiple sources are used to generate different QKD states. Here we present a QKD source design and an evaluation of critical subsystems characterized by the Quantum Bit Error Rate, Quantum State tomography, and achievable Key Rates. The QKD source is intended to operate within compact Size, Weight, and Power constraints. The Polarization-Modulator QKD source has applications in future mobile quantum networks such as Unmanned-Aerial Vehicles (UAV) and autonomous vehicles, as well as fixed fiber-based quantum networks.
Quantum mechanics can produce correlations that are stronger than classically allowed. This stronger-than-the-classical correlation is the “fuel” for quantum computing. In 1991 Schumacher forwarded a beautiful geometric approach, analogous to the well-known result of Bell, to capture the non-classicality of this correlation for a singlet state. He used well-established information distance defined on an ensemble of identically–prepared states. He calculated that for certain detector settings used to measure the entangled state, the resulting geometry violated a triangle inequality —a violation that is not possible classically. This provided novel information–based on geometric Bell inequality in terms of a “covariance distance.” Here we experimentally reproduce his construction and demonstrate a definitive violation for a Bell state of two photons based on the usual spontaneous parametric down-conversion in a paired BBO crystal. The state we produced had visibility of Vad = 0.970}0.012. We discuss generalizations to higher dimensional multipartite quantum states.

Parental sensitivity, parental intrusiveness, child involvement, and dyadic synchrony are factors associated with important developmental outcomes. The current study uses the CIB (Feldman, 1998) for a behavioral approach to investigating the effects of parental loneliness and perceived social support during the COVID-19 pandemic on parent-infant (12 􀂱 24 mo) behaviors during a free-play interaction (N = 43). The results of this study demonstrate that parental loneliness has an adverse effect on parental sensitivity and intrusiveness, but not child involvement and dyadic synchrony. Mediation analyses revealed the specific mechanisms by which low levels of perceived social support increase loneliness, which in turn, influences parents to behave more intrusively towards their infants. The findings of this study hold value for interventions that seek to reduce risk for adverse developmental outcomes. These results also hold important implications about the COVID-19 pandemic and its negative impact on parenting behaviors towards infants.