Environmental and Occupational Health Sciences Institute
Rutgers University Community RepositoryNot a member yet
58345 research outputs found
Sort by
Searching for optimal solutions in two settings
No full textThis thesis consists of two projects divided into three chapters. In the first chapter, we provide optimal allocation of units for factorial experiments under randomization-based inference. Existing optimal allocation results typically assume a super-population model and that the outcome data comes from a known family of distributions. Instead, we focus on randomization-based causal inference for the finite-population setting, which does not require model specifications for the data or sampling assumptions. We propose exact theoretical solutions for optimal allocation in factorial experiments under complete randomization with A-, D- and E-optimality criteria. We then extend this work to factorial designs with block randomization. We also derive results for optimal allocations when using cost-based constraints instead of sample size constraints. To connect our theory to practice, we provide convenient integer-constrained programming solutions using a greedy optimization approach to find integer optimal allocation solutions for both complete and block randomization. The proposed methods are demonstrated using two real-life factorial experiments conducted by social scientists. In the second chapter, we extend the results from the first chapter to fractional design settings and provide ways to choose objectively between competing fractions of a given design. Classical criteria such as maximal resolution and minimum aberration may not be able to differentiate in all cases objectively between fractions, and thus, our work is an important solution to this problem. We take practical considerations into account in our methods to give researchers multiple criteria and ways to choose between fractions. Finally, we pick a motivating example from the literature to demonstrate our results. In the third chapter, we embark on our second project, a response surface optimization problem in molecular dynamics. ReaxFF is a molecular dynamics method that can be considered a good approximation to quantum methods for investigating reactive molecular systems that consist of tens of thousands to hundreds of thousands of atoms. While ReaxFF is usually a much faster alternative to quantum methods, the force field consists of nearly 100 parameters per element, which makes the force field development a high dimensional optimization problem. In addition to the high-dimensionality, non-convexity and non-continuity make it a hard problem to optimize. We use random optimization along with coordinate search strategies to optimize efficiently and sample new parameter points that yield good molecular properties close to predefined `reference values' obtained from quantum mechanical methods for the system. We also provide empirical error guarantees starting from any random sample of inputs. In addition, we discover new points for the system at adjusted error levels of as compared to literature at levels under the same loss function, registering over 80\% improvement.Ph.D.Includes bibliographical reference
Disclosure of suicidal thoughts and behaviors: an investigation of facilitators and barriers
No full textEffective suicide risk assessment hinges on obtaining accurate information about an individual’s history of suicidal thoughts and behaviors (STBs). However, the reliability of these assessments may be compromised by at-risk individuals’ reluctance to disclose their experiences fully. The present study aimed to: 1) examine barriers to and facilitators of accurate STB disclosure; 2) evaluate the extent to which past accuracy of STB disclosure and intention to disclose future STBs are associated with current suicide risk factors (e.g., stigma, perceived burdensomeness, thwarted belongingness, hopelessness, suicide attempt history); 3) assess the association between accuracy of STB disclosure – as well as intention to disclose future STBs – and endorsement of disclosure facilitators. Overall, when participants were asked about suicide factors by most clinicians, they were not asked about firearm possession. Furthermore, only a fraction of participants reported being comfortable accurately disclosing firearm possession across clinician types. A sample of 159 adults with past STB assessment experiences by medical or mental health professionals participated in an online survey. Findings revealed that STB disclosure inaccuracies were common, with many participants portraying STBs as less intense or frequent than they actually were, and hiding desire to die by suicide. Participants reported significantly greater accuracy of STB disclosure, as well as greater intention to disclose future STBs, to psychiatrists and psychologists/therapist, compared to non-psychiatrist medical doctors and nurses. Fear of hospitalization, embarrassment, and judgment were among the most frequently endorsed barriers to suicidal ideation disclosure, while the most endorsed facilitators included desire for emotional support and mental health treatment. Inconsistent with hypothesis, past accuracy of STB disclosure was not significantly related to suicide risk indicators such as current suicide ideation severity, hopelessness, thwarted belongingness, perceived burdensomeness, acquired capability for suicide, suicide-related stigma, help-seeking stigma, negative urgency, emotion regulation difficulty or suicide attempt history, controlling for covariates. Current intention to disclose future STBs to providers was also not significantly related to these variables, with the exception of impulse control difficulties. However, past accuracy of suicidal ideation disclosure was positively associated with current intention to disclose future ideation, suggesting a role for clinicians in using past disclosure experiences to gauge the reliability of current assessments. Current intention to disclose future STBs significantly predicted endorsement of facilitators, such that individuals with greater current intention to disclose future suicidal ideation were more likely to have endorsed wanting support in seeking mental health treatment, wanting to receive medications, and being taken seriously as factors that facilitated past suicidal ideation disclosure. These findings underscore the need for targeted interventions that reduce disclosure barriers and enhance facilitators, thereby improving the accuracy of suicide risk assessments. Future research should focus on developing brief, effective strategies for healthcare providers to encourage accurate disclosure and provide appropriate interventions.Ph.D.Includes bibliographical reference
Safari: a great adventure
No full textThis thesis explores the cultural history of Jungle Habitat, a drive-through safari park in West Milford, NJ, that was owned and operated by Warner Bros. from 1972 to 1976. Working back from its dates of operation, this thesis explores three chronologies that provide insight into Jungle Habitat’s existence as a strange confluence of wild animals, animal shows/rides, Looney Tunes shows and African dance performances. These chronologies, stemming from the late nineteenth and early twentieth centuries, are: (1) American safari expeditions to Africa to collect animal specimens; (2) the rise of hyper-realistic taxidermy dioramas in natural history museums; and (3) the capture of the non-human world using motion picture technology. Ideas from Donna Haraway’s feminist piece, “Teddy Bear Patriarchy: Taxidermy in the Garden of Eden, New York City, 1908-1936” and John Berger’s essay, “Why Look at Animals?” position these chronologies within the capitalist and colonialist legacy of wildlife conservation in the United States and as imaginaries for human behavior. The actions of President Theodore Roosevelt and taxidermist Carl Akeley figure prominently in this discussion. A visual synthesis consisting of archival photographs, custom View-Master reels, a site map, timelines, and artists’ books ties these threads of history and social criticism together.M.L.A.Includes bibliographical reference
Navigating uncertainties in power system reliability and renewable energy integration: strategies and methodologies
No full textThis dissertation undertakes an extensive investigation into the reliability analysis and optimization of power network systems amidst uncertainties, emphasizing the incorporation of novel strategies in energy market operations. Contemporary power networks encounter a myriad of uncertainties, from the variability of renewable energy sources and transmission infrastructure vulnerabilities, such as line and tower failures, to unpredictable demand fluctuations. These uncertainties compromise the network's reliability and complicate its design, adversely affecting the energy market's operational efficiency. Despite significant research in this field, much of the existing literature concentrates on traditional aspects of network reliability and optimization algorithms, often overlooking critical factors relevant to current power network dynamics and energy market complexities. This dissertation aims to fill these gaps by offering a thorough analysis of power network reliability and design optimization under uncertainties, with a specific emphasis on innovative energy market operation strategies. It is divided into two main sections: (1) a detailed examination of reliability analysis and the optimal design of power networks under uncertainties, and (2) an investigation into the optimal energy scheduling in the energy market amidst uncertainties.The first section of this dissertation introduces the Eigenvector Dimension Reduction (EDR) method for reliability analysis in the face of time-dependent load uncertainties. This novel approach offers a more efficient alternative to traditional Monte Carlo simulations, enabling precise probability-based evaluations of system reliability without leading to unnecessary system expansion. The EDR method is then utilized for the reliability-based design optimization (RBDO) of distributed generators' (DGs) placement to enhance power transfer quality and minimize losses under uncertain conditions. A modified binary particle swarm optimization (B-PSO) algorithm is also introduced, showcasing its effectiveness in optimally locating DGs to improve both system reliability and operational efficiency.
The second section delves into the integration of credit-based scheduling in energy market operations, particularly targeting the intermittency issues of renewable energy sources. It presents a pioneering framework for day-ahead market scheduling that incorporates a credit score system for renewable power producers, established through Bayesian inference, to enhance both the operational efficiency and economic benefits of market transactions. Additionally, a Stackelberg game-theoretic model is introduced for strategic imbalance pricing and capacity bidding within a credit score-driven energy market, aimed at improving market efficiency and aligning with renewable energy targets. The dissertation also outlines a bidding capacity strategy for a state-of-the-art hybrid charging station for electric buses, which acts as a power supplier in the market. This strategy is crafted to navigate the challenges of renewable energy intermittency and price volatility, offering innovative methods for integrating renewable sources into the market and maintaining economic sustainability.
In summary, this dissertation makes a substantial contribution to the fields of reliability analysis and system optimization in power networks, introducing novel solutions for the integration of renewable energy sources into energy market operations. The research presented provides critical insights for policymakers, system operators, and market participants, with the goal of improving the reliability, efficiency, and sustainability of future power systems.Ph.D.Includes bibliographical reference
Two random processes for approximating solutions to combinatorial problems
No full textThis thesis concerns two problems at the intersection of discrete mathematics and theoretical computer science. Broadly speaking, both of these problems regard the outcome of a random process on a discrete structure. It is well known that any graph of maximum degree can be properly edge colored with at most colors. In the online setting, it has been a matter of some interest to find an algorithm that can properly edge color any graph on vertices with maximum degree using at most colors. Here we study the na"{i}ve random greedy algorithm, which simply chooses a legal color uniformly at random for each edge upon arrival. We show that this algorithm can -color a graph for arbitrary in two contexts: first, if the graph is fixed ahead of time and its edges arrive in a uniformly random order, and second, if the edges of the graph are selected adaptively by an adversary as the algorithm progresses, but with the restriction that . A corollary of the second result is that there must exist a deterministic algorithm to -color dense graphs.
For the second problem, we study probability distributions on a set with an associated weight function . For any such distribution, , we define a defending distribution to be a distribution on subsets of of total weight at most 1 that satisfies certain constraints with respect to both and linear orderings on . The existence of a defending distribution is relevant to the field of social choice theory, where it was shown in a paper by Jiang, Munagala, and Wang that it implies the existence of a stable lottery, a method of selecting committees that is ``fair" in some sense. We will extend a result from that paper to show that under certain restrictions of the weight function, a defending distribution always exists.Ph.D.Includes bibliographical reference
Poetics of the self: women, French cinema
No full textThis dissertation redefines the interplay between womanhood and cinema exploring the poetic dynamics inherent in their connection. Coming together like a six-part music piece, emphasizing the harmonious coexistence of multiple aspects of the feminine voices in cinema, this work addresses historical representations of women creators in French cinema and highlights how the filmmakers whose work I study have addressed gender imbalances in visual culture. It then explores how the introduction of a female gaze shifts the dynamics traditionally dominated by male perspectives and examines the expression of feminine anxieties, passions, and social realities. In this work, I further trace the political implications and authorial identities that emerge from women filmmakers’ works, suggesting that these works not only reflect on the creative act itself but also challenge entrenched patriarchal structures. This dissertation examines the terminology used to define women's roles in cinema, considering whether such terms imply a sense of alterity and how historical and emerging discourses either perpetuate or dissolve this alterity. The concept of “authorship” in cinema is then reinterpreted through the lens of femininity, providing alternative insights into its meaning. The structure of the project reflects upon, with its metaphor of a choral composition, the personal experiences and stories told by women in film, resonating as a soprano voice. The mezzo-soprano voice reflects on the act of narration and fiction-making and offers new perspectives on the artist figure. Finally, the alto voice compares the political aspects of women’s cinematic expression to a foundational line that supports the personal and reflective dimensions, highlighting a drive toward emancipatory expressions within social structures and moral dynamism.Ph.D.Includes bibliographical reference
Blood Wastage in the Cardiothoracic Intensive Care Unit
No full textPurpose of Project: The aim of this quality improvement project was to reduce blood product wastage in the cardiothoracic intensive care unit (CTICU) by incorporating a revised handoff sheet that included a reminder prompt as a new standard clinical practice. The prompt reinforced nurses to properly store blood products in the unit’s refrigerator or return to the blood bank. Methodology: The CTICU’s aggregate blood wastage from 2020 to May 2023 was collected as baseline data before the revised handoff sheet was implemented over a 12-week period. Handoff sheets were then collected and analyzed, as well as monthly data reports of product wastage.
Results: Wastage events marginally decreased after implementation of the revised OR handoff sheet. There were six events in the pre-intervention and one event in the post intervention. Additionally, there was 90% compliance on dayshift and 83% on nightshift in completion of the handoff sheet.
Implications for practice: Use of a reminder prompt on an OR handoff sheet is a significant QI initiative in blood waste management.D.N.P.Includes bibliographical referencesIncludes vit
China and Russia versus the United States of America: geopolitical competition in the United Nations Security Council.
No full text1). Rationale for the study.
The international system is transforming and headed towards a multipolar order. The coercive factor responsible for this ongoing transition is the geopolitical competition between America and a rising China. Beijing is encouraged by a resurgent and radical Moscow. The two emerging Powers show alignment during international relations. The rapprochement produces tensions that challenge American hegemony. Literature on the alignment is nuanced but decipherable in scholars recognizing the phenomenon’s potency and others downplaying it. The Dissertation’s innovative approach introduces a new way of investigating the alignment. It examines China’s and Russia’s unique Security Council veto and non-veto behaviors towards America. Its profound conclusions enhance our knowledge and understanding of the Sino-Russo alignment.
2). Scope.
The research covers twenty years of the new millennium, from 1/1/2001 to 12/31/2020.
3). Research Question.
Do China’s and Russia’s veto and non-veto signatures systematically corroborate both actors’ alignment against America?
4). Hypotheses.
China’s alignment with Russia lacks evidential support from their United Nations Security Council encounters.
5). Argument. China and Russia are not necessarily aligned against the United States at the Security Council. China’s behavior towards America suggests a rational actor, while Russia displays the Great Power attitude.
6). Method.
Quantitative and qualitative surveys were conducted at the United Nations Security Council, with the investigation focusing on 38 case studies. The survey was extended to the American White House, various think tank organizations, the American Department of State, and the Foreign Affairs Ministries of China and Russia. Equally, several open media sources, alongside some bilateral and multilateral agreements involving China, Russia, and America, were scrutinized.
7). Results.
Distinct differences exist in the three actors' veto and non-veto behaviors. Moreover, the premise behind China or Russia’s use of the veto varies.
8). Conclusion.
An analysis of the unique veto and non-veto encounters among the three actors at the Security Council does not support China’s alignment with Russia to challenge American hegemony.Ph.D.Includes bibliographical reference
Molecular design of organic/inorganic hybrids for solar energy conversion
No full textThis thesis comprises four chapters and three distinct research projects. Chapter 1 of this dissertation provides an introduction and background on solar energy conversion and storage systems, such as Dye-Sensitized Photoelectrosynthesis Cells (DSPECs) and Dye-Sensitized Solar Cells (DSSCs). It offers an overview of the components and application of dye molecules for sensitizing metal oxide (MOx) electrodes in DSSCs and DSPECs. Furthermore, it provides details on MOx electrodes and electron transfer processes occurring in these devices, which can potentially enhance or reduce their efficiency. Chapter 2 describes the first project, in which we explored how the structure of anchor groups forming covalent bonds to nanostructured semiconductors influences the binding of light harvesters to mesoporous TiO2 thin films. Specifically, we introduced a third coordination unit (X) into isophthalic acid derivatives (2-hydroxyisophthalic acid, 2,6-dicarboxypyridine, and 2,6-dicarboxypyridin-1-ium) and studied their effects, along with 1-ethynylpyrene as a chromophore, both in solution and on mesoporous TiO2 thin films.The introduction of X at C2 of isophthalic acid had significant implications for Highest Occupied Molecule Orbital (HOMO) and Lowest Unoccupied Molecular Orbital (LUMO) energies, orbital’s electron distributions, and binding strength to TiO2 films.Chapter 3 describes the second project. In this project, a family of ruthenium-based rigid rods were anchored to mesoporous Indium Tin Oxide (ITO) nanocrystal films. These compounds featured different numbers of conjugated phenylacetylene units in the bridge that is between the ruthenium complex and a 1,3-dicarboxylate isophthalic anchoring group. Our investigations revealed that the presence of phenylacetylene bridge units had minimal impact on redox properties and visible light absorption spectra, but greatly influenced injection rates. The rate of electron transfer from ITO to the rigid rods was measured by tuning the Gibbs free energy change (-ΔG°), with pseudo-rate constants increasing as the length of bridge decreases. These findings provided insights into reorganization energy (λ) and electronic coupling matrix element (Hab) variations and their potential impact on solar energy conversion applications. This project was conducted in collaboration with Dr. Gerard J. Meyer’s group at the University of North Carolina-Chapel Hill. The focus of this thesis is on the synthesis of ruthenium complexes used in this study. The kinetic data were collected and analyzed by Quentin Loague in Dr. Meyer’s group at University of North Carolina-Chapel Hill and are briefly mentioned in this thesis.
Chapter 4 summarizes the third project, which explores the influence of Self-Assembled Monolayers (SAMs) formed from disulfide lipoic acid-based compounds on Ag electrodes to improve the durability and efficiency of optoelectronic devices. These compounds offer double binding sites, potentially enhancing device performance compared to conventional thiol compounds that attach to the surface through a single site. By fabricating inverted organic semiconducting hole-only devices, we studied the influence of SAM structure on charge injection in the presence of (±)-α-lipoic acid, (±)-4-phenylbutyl 5-(1,2-dithiolan-3-yl) pentanoate, and iso-lipoic acid, which is an achiral form of lipoic acid. These SAMs can potentially enhance charge injection by altering the Ag work function or modifying the interaction between the metal surface and the polymer. This research details how SAM functional groups and bonding interactions influence the performance of organic semiconductor devices. Device performance studies were conducted in Dr. Deidre M. O’Carroll’s group at Rutgers University-New Brunswick.
These projects contribute to the understanding of various aspects of Interfacial Electron Transfer (IET), metal oxide/metal electrodes, molecular design, and device optimization for energy conversion and storage applications.Ph.D.Includes bibliographical reference
Geometry, Interpretability, and security of neural networks
No full textThis work addresses two pivotal challenges in deep learning: ensemble methods and Neural Network (NN) security. We introduce methods and concepts that advance both fields and explore crucial connections between them. Our work on Geometric Ensembles provides a leap forward in Neural Machine Translation (NMT) technology and offers a fresh perspective on standard techniques that use machines to translate human language. Unlike traditional methods that require complex interactions with the underlying systems, Geometric Ensembles work without any need for the details of the underlying system, instead working directly on output sentences, making them black-box. Our comprehensive empirical studies across three standard MT datasets demonstrate that Geometric Ensembles not only rival the accuracy of standard methods but also provide significant gains in efficiency and interpretability. We confront the crucial problem of NN security with the introduction of Secure Engravings, a method to embed secret (input, output) information within an NN such that they remain undetectable without knowledge of the secrets themselves. In an era where NNs are increasingly being used in critical decisions, this work on trust and security in training these systems is of paramount importance. This study provides a comprehensive framework that includes (1) definitions of security for Secure Engravings, (2) Machine Learning (ML) methods for constructing these networks, and (3) a threat model instantiated with advanced interpretability methods to devise distinguishers/attackers. We demonstrate that while some engravings are vulnerable, our more secure constructions withstand attacks across multiple datasets. Secure Engravings also offer a novel benchmark for evaluating interpretability methods, a contribution of independent significance to the field. Finally, we bridge these two domains by proposing Securely Engraved Ensembles--sharing among each base system different subsets of the special (input,output) pairs allowing them--as a collective--the ability to realize the entire set. We derive theoretical bounds on the information distribution and provide heuristic insights into how these ensembles can enhance security potential. Our proof-of-concept on an image classification task demonstrates the potential of Securely Engraved Ensembles.Ph.D.Includes bibliographical reference