Treasures @ UT Dallas
Not a member yet
7697 research outputs found
Sort by
Loneliness in the United States: a Sociological Approach
No full textLoneliness in American society has received much attention in recent years from researchers and
public health officials. Loneliness is defined as a form of perceived detachment from others
which causes emotional distress, while social isolation is an objective measure of the degree of
social contact an individual has. Loneliness and the related concept of social isolation are linked
to various detrimental health outcomes, with some researchers viewing them as a threat to public
health akin to smoking or obesity. This paper aims to examine the current landscape of loneliness
within the United States, reviewing how it manifests within the various diverse groups which
make up American society, and how the recent COVID-19 pandemic has affected social
connectedness. Research within the field of sociology is discussed to illustrate the value of
applying a sociological approach to loneliness and social isolation, both in understanding the
origins of these social ills and in devising solutions aimed at reducing loneliness and social
isolation within American society
Sensitizing Conventional Drug Vehicles to Ultrasound for Improved Delivery
No full textA successful drug delivery system can be defined as a platform that enables the introduction of the
optimal drug dose to a specific tissue for an intended duration. As the definition suggests, rigorous
control over drug deposition is a necessity. Drug-based therapy historically included the
introduction of a drug to the bloodstream, mainly using the digestion route or injection. This
traditional delivery method has no specificity or control over delivery to the diseased tissue.
Furthermore, administration of the free drug in the bloodstream results in a narrow therapeutic
window where only small doses of drugs are used to prevent detrimental drug bioeffects. One of
the significant accomplishments in pharmaceutics was designing drug carriers that can encapsulate
high payloads of medicines and reduce drug toxicities in the bloodstream. This accomplishment
was mainly a fruit of cancer research where encapsulation of highly toxic chemotherapeutics is
critical.
In the 1990s and 2000s, drug release from carriers and successful drug targeting were major
research areas. One of the promising drug delivery platforms introduced in that era was image-
guided drug delivery (IGDD). IGDD is a platform that uses a type of external energy source to
activate and control drug release from stimuli-sensitive drug carriers. One of the critical choices
in IGDD platforms is selecting a suitable external energy source. Ultrasound (US) has been one of
the primary modalities for IGDD since it is spatiotemporally precise and cost-effective.
Despite successful preclinical and clinical data in the ultrasound-mediated drug delivery field,
there is still no clinically approved ultrasound-sensitive drug carrier. In this thesis, we report novel
methods to make conventional drug delivery vehicles sensitive to ultrasound for producing safe
and efficient ultrasound-responsive drug vehicles. We chose clinically successful drug vehicles for
this purpose and delved deeper into their material chemistry to manipulate and sensitize them to
ultrasound energy. We hypothesized that by introducing ultrasound-sensitive agents to traditional
drug vehicles, we could build ultrasound-sensitive drug carriers suitable for drug delivery and
imaging. The designed drug carriers have the potential to be used in image-guided drug delivery
platforms. Our results suggest that we can create ultrasound-sensitive particles and agents suitable
for theranostic/dual imaging applications by incorporating ultrasound cavitation nuclei in a drug
delivery system
Necessary Evils: the Role of Horror in Modern and Contemporary Literature
No full textIn my extensive studies of horror, I have found that the genre of horror has typically not been
taken seriously in its own right. This can be extended to the occurrence of horror in other types
of literature. Horror, if recognized at all, is viewed as a component of the story, and not
necessarily as significant or as relevant as other aspects of the book. This dissertation
approaches the problem of how literary horror can, genre or otherwise, be recognized as a part of
legitimate and influential academic study and why such study is important. To do so,
I examine multiple works of both genre horror and literary horror, using established literary
theories to analyze and understand these written works. I also examine multiple works not
classified as horror yet contain instances of significant horror to show that horror exists past the
genre. I utilize literary theories such as the uncanny, the monstrous, Kristeva’s theory of
abjection and the Jungian shadow to show the literary merits of these works.
To not read horror is to ignore aspects of life that act as a mirror reflecting society and individual
fears at any point in time. Such willful evasion can be detrimental. Dismissing horror as merely
entertainment avoids the social and cultural deceits it can expose. Horror is an interpretation of
what is both desired and feared in our lives. Its omnipresence makes it critical to be understood
as a vehicle used to acknowledge and understand our fears, and ultimately determine the best
way to handle them
Flows With Geometric and Topological Constraints
No full textWe describe several results on flow problems with geometric and topological constraints.
More specifically, we present two near-linear time approximation schemes for the geometric
transportation problem. The first one is a randomized algorithm running in near-linear time
for arbitrary input, and the second one is a deterministic algorithm running in near-linear
time if the input has a polynomially bounded spread. While these two algorithms solve the
transportation problem with certain geometric constraints, in the rest of this dissertation
we describe some new preliminary observations for computing maximum flows in planar
embedded graphs as well as some new algorithms inspired by those observations.
We first describe an algorithm that runs in nε−O(d) logO(d) n time and returns a (1 + ε)-
approximation to the optimal transportation map for the geometric transportation problem, with high probability. This algorithm is the first one that runs in near-linear time
for any input. It first reduces the problem into the problem of seeking a minimum cost
flow of a sparse spanner with a (1 + O(ε))-distortion of the point-to-point distance, then
computes a (1 + O(ε))-approximation to the minimum cost flow using Sherman’s preconditioning framework, and finally extracts a transportation map from the minimum cost
flow without increasing the cost. The second algorithm for the geometric transportation problem we present is a deterministic (1 + O(ε))-approximation scheme running in
nε−O(d)(log n + log2 Sp(P ) log log Sp(P )) log Sp(P ) time where Sp(P ) represents the spread
of the point set P given by the input. It uses a spanner with a deterministic (1 + O(ε))-
distortion by increasing the size of the spanner by a factor of at most 2d in comparison to
the one we use in the first algorithm.
We close by describing some new results for computing maximum flows in planar embedded
graphs. We describe a novel way to analyze the number of times a dart u → v can enter the
residual (u, t)-path when we compute maximum flows using leftmost pushes. Eisenstat and
Klein describe a linear-time algorithm for a special case of the maximum st flow problem
where the arc capacities are small integers. Our new observations show their algorithm could
actually run in O(n2) time in the worst case and imply an easy fix to a faulty claim of theirs.
Then we describe a new preflow-push lemma for computing maximum flows in the plane
and ongoing work in designing new maximum flow algorithms using this lemma. Finally, we
consider a special case of the multiple-source multiple-sink maximum flow problem in the
plane where all sources and sinks are among the same face. We describe the first linear-time
algorithm for this problem if all edges have small integer capacities. For the same problem, we
also describe a simple algorithm that runs in O(n log n log k) time without the small integer
capacity assumption
A Musical Conversation: Interaction in Flamenco Jazz
No full textIn the field of jazz, the intricate interplay and dynamic exchange between musicians (musical
interaction) has been well-documented. Interaction in flamenco jazz, however, has had little
scholarly discourse. This dissertation fills that gap by addressing the way musicians in the fields
of flamenco and jazz have interacted to produce a new style called flamenco jazz, which has
achieved recognition in its own right, independently from either flamenco or jazz. The primary
aim of this research is to trace the evolution of flamenco jazz, charting its progression over
centuries from a niche folkloric musical tradition to its current status as a globally recognized
musical art form integrated into mainstream culture. I also seek to illuminate how musical
interaction manifests within the context of flamenco jazz, emphasizing the significance of both
stylistic and interpersonal dynamics in its development. This study draws upon philosophical
frameworks taken from the field of semiotics, historical references, interviews with notable
flamenco jazz musicians, and in-depth analysis of flamenco jazz recordings to illustrate the
nuances of musical interaction within the genre. By examining the interplay of melodic motifs,
harmonic structures, and rhythmic patterns, my research elucidates the intricate relationships that
underpin flamenco jazz performance. Concluding my study, I reflect upon the pivotal role of
musical interaction in the creative process, juxtaposing its significance within the contexts of
African American and Spanish cultures. I argue that there exists untapped potential for further
advancement within the flamenco jazz genre, asserting its ability to serve as a bridge between
diverse cultures and generations
Towards Functional Safety in Deep Learning Hardware Accelerators
No full textThis dissertation furnishes an extensive end-to-end approach to improve the functional safety
of Deep Neural Network (DNN) accelerators. Deep Learning (DL) algorithms, with their
gamut of efficient image and video processing capabilities, have emerged as promising can-
didates to be deployed across low-latency mission-critical systems at the edge. Ongoing
research to obtain high-performance DNN executions has led to the development of several
purpose-built low-overhead inference hardware accelerators. However, such DNN acceler-
ators are susceptible to hardware faults, due to high-energy particles, process variations,
temperature and structural deformities manifesting as latent defects. These faults can intro-
duce misclassifcation, thereby jeopardizing the Functional Safety (FuSa) in mission mode,
which can eventuate to disastrous consequences, including loss of human lives. Since the de-
cisions ascertained by a DNN accelerator is often beyond human control (e.g., a surveillance
drone executing computer vision responsibilities), FuSa assurance of such inference accel-
erators is extremely critical. This research addresses this crucial problem by developing a
comprehensive lifetime FuSa maintenance framework, encompassing FuSa assessment, FuSa
violation detection and FuSa assurance for resource constrained environments.
Although existing research have analyzed the impact of manufacturing faults in such in-
ference hardware with respect to yield improvement, extensive FuSa assessment and FuSa
maintenance schemes for in-field DNN accelerators in operational or mission mode have been
largely unexplored. This dissertation, for the first time to our knowledge, provides a scientific
foundation for functional safety of DNN accelerators. The research to develop a four-stage
FuSa improvement framework leads to the following outcomes: 1) furnishes a comprehensive
assessment on the dependency of FuSa in a DNN accelerator on multivariate fault character-
istics and underlying DNN architectures, 2) generates optimal set of functional test patterns
to be employed to detect FuSa violation dynamically, at runtime, 3) diagnoses the locations
of manifested hardware faults to improve FuSa assurance using diagnostic test patterns and a
hardware-DL combined diagnostic algorithm, and 4) develops software and hardware-based
mitigation strategies in the context of manifested faults, obviating the need for retraining
the DNN or replacing the accelerator.
To conclude, this dissertation, theoretically, helps bridge the gap between low-level electrical
faults and high-level DNN misclassification during inference. In practice, this dissertation
facilitates the designing of functionally safe and robust DNN inference accelerators that
are used in safety-critical executions in mission mode, like, autonomous driving, unmanned
aerial vehicles, smart healthcare applications, industrial robots, and internet of military and
battlefield things
Stability and Bifurcation Analysis of a Delay Differential Equation Modeling the Human Respiratory System
No full textThe human respiratory system takes oxygen and releases carbon dioxide in air with the
help of lungs and other respiratory organs. Understanding the human respiratory system is
important for many medical conditions. There are many disorders associated with respiratory system that affect a large number of people. Thus studying this system with a good
mathematical model has far-reaching implications.
We study the two state model which describes the balance equation for carbon dioxide and
oxygen. These are nonlinear parameter dependent and because of the transport delay in
the respiratory control system, they are modeled with delay differential equation. So the
dynamics of a two state one delay model are investigated. By choosing the delay as a
parameter, the stability and Hopf bifurcation conditions are obtained. We notice that as the
delay passes through its critical value, the positive equilibrium loses its stability and Hopf
bifurcation occurs. The stable region of the system with delay against the other parameters
and bifurcation diagrams are also plotted. The three dimensional stability chart of the two
state model is constructed. We find that the delay parameter has effect on the stability but
not on the equilibrium state. The explicit derivation of the direction of Hopf bifurcation and
the stability of the bifurcation periodic solutions are determined with the help of normal
form theory and center manifold theorem to delay differential equations.
We also study the five state model with four delays numerically. We investigated the stability
of the system at several delays and stability chart near the measured values are constructed.
Finally, some numerical example and simulations are carried out to confirm the analytical
findings. The numerical simulations verify the theoretical results
Locating the South Through American Literary Conversations
No full textMuch of the scholarly discussions and the pedagogical representations of American literature
have placed undue regional boundaries as they read American literature as a mere regional
cultural production. This dissertation offers a long-needed alternative to that approach by
highlighting the responsive nature of the American literary canon, where authors from different
regions negotiate the overall American story. In an introduction and four chapters, this
dissertation draws upon the theoretical framework and consults some of the most recent and
exciting ideas emerging from the “new southern studies” movement. The first chapter traces
literary representations of colonial America by analyzing texts on President Thomas Jefferson’s
affair with his slave mistress, Sally Hemings, which resulted in the birth of six children,
specifically how Barbara Chase-Riboud’s Sally Hemings engages William Wells Brown’s
Clotel; or, the President’s Daughter. The second chapter reads two major slave narratives:
Frederick Douglass’ Narrative of the Life of Frederick Douglass and Miguel Barnet’s Biography
of a Runaway Slave, as well as a neo-slave narrative novel, Toni Morrison’s Beloved, to unfold
the myth of freedom often associated with the genre. The third chapter reads Richard Wright’s
Uncle Tom’s Children as a direct response to Harriet Beecher Stowe’s Uncle Tom’s Cabin,
rejecting the latter’s commitment to moral suasion through the assertion of violence as the
ultimate drive to freedom. The final chapter reads William Faulkner’s inaugural novel, Soldiers’
Pay, in between F Scott Fitzgerald’s The Great Gatsby and Ernest Hemingway’s The Sun Also
Rises, with the rise of the new woman figure as a thread connecting the three writers. This
dissertation concludes by asserting the need to reassess the way we think about American
literature as one continuous narration with a single unified literary canon
Role of FimK in Mediating Host Urinary Bladder Epithelial Cell Association of Uropathogenic Klebsiella Pneumoniae and Quasipneumoniae
No full textKlebsiella spp. commonly cause both uncomplicated urinary tract infection (UTI) and
recurrent UTI (rUTI). K. quasipneumoniae, a relatively newly defined species of
Klebsiella, has been shown to be metabolically distinct from K. pneumoniae, but its
urovirulence mechanisms have not been defined. Type 1 and type 3 fimbriae, encoded
by fim and mrk operons respectively, mediate attachment of Klebsiella spp. to host
epithelial cells. fimK is a regulatory gene unique to the Klebsiella fim operon that encodes
an N-terminal DNA binding domain and a C-terminal phosphodiesterase domain that has
been hypothesized to cross-regulate type 3 fimbriae via modulation of cellular levels of
cyclic di-GMP. Comparative genomic analysis between K. pneumoniae and K.
quasipneumoniae revealed a conserved premature stop codon in K. quasipneumoniae
fimK that results in loss of the C-terminal phosphodiesterase domain (PDE). We
hypothesized that this truncation would ablate cross-regulation of type 3 fimbriae in K.
quasipneumoniae. Here, we report that K. quasipneumoniae KqPF9 bladder epithelial
cell association and invasion is dependent on type 3 but not type 1 fimbriae. Further, we show that basal expression of both type 1 and type 3 fimbrial operons as well as bladder
epithelial cell association are higher in KqPF9 than in K. pneumoniae TOP52.
Interestingly, complementation of KqPF9∆fimK with the TOP52 fimK allele markedly
reduced type 3 fimbrial expression and bladder epithelial cell attachment, a phenotype
that was rescued by mutation of the C-terminal PDE active site. Taken together these
data suggest that C-terminal PDE of FimK modulates type 3 fimbrial expression in K.
pneumoniae and its absence in K. quasipneumoniae leads to a loss of type 3 fimbrial
cross-regulation
Addressing Testing and Verification Challenges in a Transistor-level Programming Fabric
No full textA TRAnsistor-level Programming (TRAP) fabric was developed for integrated circuit redac-
tion to safeguard hardware intellectual property from untrusted foundries. By pushing pro-
grammability to the transistor level, TRAP drastically reduces the area overhead while at
the same time introducing more formidable challenges for both brute-force and intelligent
search-based attacks. Adoption of this novel fabric necessitates an RTL-to-GDSII tool flow,
including a verification solution and a method for testing the chips for manufacturing defects.
Bidirectionally-operated primitives are required to represent the granular transistor-level ar-
chitecture of TRAP. Existing formal verification and pattern generation techniques do not
support the primitives necessary to model such transistors in hardware description languages
(HDL). Consequently, verifying and testing TRAP using traditional methods is not inher-
ently feasible.
A method to accurately model TRAP for simulation-based verification is proposed. The
solution is extended to model a programmed TRAP fabric by implementing the extracted
directionality for a Logic Equivalence Checking (LEC)-based verification solution. Addi-
tionally, an Automatic Test Pattern Generation (ATPG)-based testing solution is proposed
for transistor-level fabrics, and finally, a design aware TRAP-specific testing solution is pre-
sented. Studies were conducted to evaluate the effectiveness of the proposed solution. The
scalability of the LEC-based verification solution is examined concerning the growing fabric
size, while the ATPG-based test solution is assessed based on the test application time of
generated configurations