Treasures @ UT Dallas
Not a member yet
7697 research outputs found
Sort by
Task Attributes, Technological Change and the Vulnerability of Employment: a State Level Investigation
No full textTwo distinct trends can prove the existence of technological unemployment in the US. First,
there are more open jobs than the number of unemployed persons looking for a job, and
second, the shift of the Beveridge curve. There have been many attempts to find the cause
of technological unemployment. However, all of these approaches fail when evaluating the
impact of modern technologies on the employment future. This study hypothesizes that
rather than looking into skill requirement or routine non-routine discrimination of tasks, a
holistic approach is required to predict which occupations are going to be vulnerable with the
advent of this 4th industrial revolution, i.e., widespread application of AI, ML algorithms,
and Robotics. Three critical attributes are considered: bottleneck, hazardous, and routine.
Forty-five relevant attributes are chosen from the O*NET database to define these three
types of tasks. Performing Principal Axis Factor Analysis and K-medoid clustering, the study
discovers a list of 407 vulnerable occupations.
The study further analyzes the last ten years (2010 to 2019) of national employment data
and finds that the growth of vulnerable occupations is only half than that of non-vulnerable
ones despite the long rally of economic expansion.
When it comes to state-level impact, the study detects a considerable disparity in job
vulnerability. While all the states will experience net job loss, some states are far more
vulnerable than others, especially the states that depend on manufacturing and production.
On the contrary, education, health care, and IT industries would enjoy hefty growth in the
coming years. The study recommends tailor-made industry and education policy for each
state based on its current level of exposure to automation. Simultaneously, it touches federal
intervention required for the betterment of underemployed and unemployed workers who
would be impacted due to technological unemployment
Taiwan and the Recognition Decisions of Developing States
No full textThis dissertation explores the following question: why do certain developing states choose to
recognize Taiwan while others do not? Developing states, despite being generally regarded as
less influential in global affairs, have been the subject of an intense diplomatic competition
between China and Taiwan. This research explores why states choose to engage with Taiwan as
well as why some states pursue informal relations rather than formal relations. A causal theory is
developed to account for how coalitions between Finance, Industry, and Labor influence the
recognition decisions of developing states. I hypothesize that countries where Finance is strongly
represented are more likely to recognize Taiwan while countries where Industry is strongly
represented are less likely to recognize Taiwan. I find quantitative and qualitative evidence to
support the theory that, in the context of informal recognition, coalitions between economic
interest groups influence the likelihood that a developing country will recognize Taiwan
Fault-proneness Prediction Based on Analysis of Human Aspects in Software Developing Process
No full textNowadays, Quality Assurance and debugging defective software is becoming gradually costly and
time-consuming. As a result, applying techniques such as fault-proneness prediction can help in
this regard in large scale software system development. The specific fault-proneness prediction
usually starts with software metrics, which are introduced to quantify and evaluated different
aspects of attributes in software processes and productions. Once we have software metrics as
indicators, a fault-proneness prediction model can be built using statistics or machine learning
methods. However, since software programs are pure cognitive products of human developers; the
flaws in it are caused by erroneous behaviors of human involved. Therefore, we feel the importance
of introducing quantitative analysis on the human factors to enhance fault-proneness prediction.
With that in mind, we come up with two approaches to achieve more precise prediction: firstly,
separating the consistent characteristics of human individuals and evaluate their future and
classifying certain working activities, and secondly, interaction during development process that
could affect the performance of developers. For the first approach, we proposed a new metric based
on historical activities: the Developer Risk Score, which is a performance indicator for the
developers’ history of making mistakes during their working period on related software projects.
Different from previous approaches, the Developer Risk Score further takes software complexity
and the severity of bug into account when evaluating the performance of a developer. Our approach
proved that for a software module that more high-risk developers involved more tend to be faultprone, and it is more efficient than other approaches when predicting such fault in software
modules. For the second approach, we considered a software network with human aspects.
Software network is a graphical model that constructed by several elements in modern software
systems to symbolize the patterns of activities between them. To integrate human aspects into the
model, we proposed the Composite Developer-Module Network. The network combines
developers’ network and software modules’ network. By using network structure categorization
comparation on the sub-structures within the network, deeper and indirect dependencies between
developers and software modules can be included. Our evaluation proved that the more complex
sub-structures of software network, that is, more developers or more function calls include in the
sub-structures, can be more correlated to possible bug introduction
Citation Release, Decarceration, and Crime in Washington, DC
No full textIn March 2020, the Washington, DC, Metropolitan Police Department expanded its policeinitiated citation release to allow officers to release subjects arrested for certain non-violent
felony offenses (ex: larceny-theft). This decarceration effort was designed to reduce COVID-19
transmission in jails and avoid maintaining custody of people pre-trial, as too many custodial
arrests would impair the operations of the Superior Court of DC during the public health
emergency. Using crime incident, arrest report, and jail population data from DC for 2013
through 2020, this dissertation investigates the effect of the citation release policy modification
(i.e., jail decarceration) and arrests on four types of economic crime: robbery, burglary, theft
from motor vehicles, and other theft. Vector autoregression analyses suggest arrests do not deter
crime and there was no detectable “decarceration” effect from expanding citation release
eligibility during the study period. Findings do not support macro-level deterrence theory or the
premise of a decarceration effect that has been identified in studies of prisons
Real-time Assessment of Obstructive Sleep Apnea Using Deep Learning
No full textSleep quality assessments provide various measures to gauge the severity of Sleep Apnea. In the
present, sleep quality testing is inconvenient for the patients in terms of both money and a
comfortable environment. Evaluation methods like the Polysomnography test require many
sensing resources. Our research proposes an inexpensive and an automated system based on
Single-lead Electrocardiogram (ECG) signal and a one-dimensional Convolutional Neural
Network classifier (CNN). We use only a single-channel ECG to measure the heart signal and
deliver them to an 1D-CNN to classify for apneic events. This method provides an alternative to
the cumbersome and expensive Polysomnography (PSG) and scoring by Rechtschaffen and Kales
visual method. In addition to this, we propose an Android application that uses a Deep Neural
Network model that we have trained to use in real assessment of Obstructive Sleep Apnea
Dynamic Resource Coordination towards Reliable and Flexible Network Slicing
No full textThe increasing demand for a diverse array of network applications entails a more flexible
and reliable networking paradigm. Network slicing is envisioned to package a set of networking, computing, and storage resources in a coordinated manner, so that the network
slice with the tailored set of resources satisfies service requirements unique to each network
application. A key enabling technology of network slicing is network softwarization, exemplified by Software Defined Networking (SDN) and Network Function Virtualization (NFV),
which enable swift migration of both networking and computing resources. While network
slicing sets a conceptual foundation for next-generation networking for diverse applications,
the resource coordination mechanism that dynamically operates and manages the resources
remains a challenging issue. This is more so when considering the computational complexity
induced by the dependency relationship among the softwarized resources and the uncertainty
of future network states, such as network failure scenarios and traffic patterns.
This dissertation features four resource allocation problems that collectively facilitate the
agile operation and management of end-to-end network slices. The first two problems are
related to the reliability aspect of network slicing. In particular, we discuss protection
and recovery problems of interdependent network components from a resource allocation
standpoint. The third problem deals with a dynamic bandwidth allocation in optical access
networks. The dynamic adjustment of allocated bandwidth assists more effective resource
utilization and the accommodation of different types of network services. Furthermore, SDN
controller placement, which determines responsiveness to a request for end-to-end resource
coordination, is examined as the fourth problem.
The theoretical analysis and proposed algorithms for the problems not only solve the specific
resource allocation tasks, but also provide fundamental insights to tackle similar allocation
problems under entangled dependency and future uncertainty. In particular, the proposed
learning-based approaches project an automated resource coordination system for more effective utilization of network resources in 5G and beyond networks
The Public as Corporate Stakeholder: Evidence from Toxic Release and Financial Reporting
No full textThis paper aims to show the role of the public as corporate stakeholder by examining the financial
reporting of polluting firms. Because pollution damages the living environment of the public, the
public “invests”, though passively, their health and life quality into polluting firms and thus
essentially is a stakeholder of polluting firms. I argue that polluting firms have incentives to report
lower profits to reduce the cost related to the pressure from the public over environmental issues.
Using corporate toxic release data, I find that polluting firms are more likely to engage in incomedecreasing earnings management when their toxic release increases. Importantly, the effect is
stronger for toxic release produced by plants located in states where residents are more likely to
pressure firms for lower pollution, suggesting that the public plays an important role. Further, the
effect is stronger for toxic release subject to stricter regulatory monitoring, for firms with higher
media coverage, and for firms in consumer product industries. However, the public still plays a
role after the influences of regulators, the media, and customers are controlled for
Evaluation and Integration of Graphene Field Effect Devices
No full textGraphene field effect devices have potential applications in emerging analogue and sensing
technology. However, to use them in end applications, evaluation of such devices in terms of
mobility, contact resistivity, and sheet resistance is pivotal. These fundamental parameters will
dictate and enable the designing of futuristic graphene-based devices. Finally, these devices should
be packaged depending on the final application or for further evaluation in an industrial
perspective. This work explores the evaluation of graphene field effect devices that has been
fabricated using Chemical Vapor Deposition (CVD) graphene source with optimization of the
device fabrication process flows on 90 nmSiO2 using materials characterization techniques such
as Raman spectroscopy, Atomic Force Microscopy, X-ray photoelectron spectroscopy, and
spectroscopic ellipsometry. The issues arising with dual-gated graphene transistor is identified and
one of the potential solutions to downscale the back-gate dielectric is demonstrated. Critical device
parameters like mobility, contact resistivity of graphene devices are evaluated and final integration
of such devices at a package level is demonstrated. For commercialization of graphene
nanoelectronics, heterogeneous integration of graphene devices on commercially available CMOS
substrate is the key and process flow for such devices is demonstrated. Additionally, a novel device
architecture to electrically dope graphene in the contact regions is identified and this could pave
way for implementation of futuristic graphene devices with tailored device properties
Seismic Data Reconstruction With Low-rank Tensor Optimization
No full textSeismic data recorded in the field often has gaps due to missing or failed receivers or aperture
restrictions and may be contaminated by external noise. Reconstruction is the process of
completing missing data and removing noise. Multi-dimensional seismic data, for example
in 3, 4, or 5D, can be efficiently stored in a tensor or multi-dimensional array. Low-rank
tensor optimization is a model used to reconstruct tensor data under the assumption that the
underlying data has low rank. Data has low rank when it has redundant rows or columns,
causing the singular values to decay at a rapid rate. Because minimizing rank is NP-hard,
a relaxation of rank can be used such as the tensor nuclear norm (TNN), derived from the
tensor singular value decomposition (tSVD). The alternating direction method of multipliers
(ADMM) effectively solves the TNN model in which the sum of singular values is minimized.
ADMM splits the minimization problem into smaller subproblems. The combination of the
ADMM method and TNN model is referred to as TNN-ADMM and is useful for reconstruc-
tion of missing data.
Exploiting the conjugate symmetry of the multi-dimensional Fourier transform (the most
expensive part of the tSVD algorithm) reduces the runtime of the tSVD algorithm for real-
valued order-p tensors by approximately 50%. The relation between the tSVD of a tensor
and the SVD of a corresponding block-diagonal matrix reveals how the singular values of the
tensor and matrix change as the orientation of the tensor changes and provides evidence for
the success of the most-square orientation when used for low rank data reconstruction. For
seismic data, the most-square orientation has frontal faces formed over the spatial dimen-
sions, so the tensor contains more redundancies than pairing a spatial dimension with time.
On real data reconstruction examples, TNN-ADMM outperforms two other data completion
methods, projection onto convex sets and multi-channel singular spectrum analysis (MSSA),
with less error and 10-1000× faster runtime.
In exploration seismology, an initial baseline survey informs decisions to produce a region,
and monitor surveys conducted during production provide updated subsurface information.
The time-lapse difference between the baseline and monitor surveys reveals changes in the
Earth due to production. Non-repeatability issues, such as inconsistent receiver locations,
negatively impact one’s ability to accurately identify time-lapse changes. If receiver locations
are regularized onto a common grid, then the baseline and monitor surveys can be compared.
The resulting tensors are incomplete due to the potential for grid blocks to not contain
receivers, so we apply TNN-ADMM to each data tensor to successfully fill in missing data,
and a time-lapse difference can be computed between the reconstructed tensors.
The unconstrained formulation of TNN-ADMM simultaneously completes and denoises data.
The convergence analysis of this method proves that the iterative solution converges to a
local minimum, provided that the step size parameter is greater than one. The convergence
proof requires new properties of the Frobenius norm for tensors, as well as Hadamard (entry-
wise) product properties. On a synthetic problem unconstrained TNN-ADMM outperforms
MSSA with 18%-27% less error and 10× faster runtime
Existence and Spatio-temporal Patterns of Periodic Solutions to Non-autonomous Second Order Equivariant Delayed Systems
No full textIn this dissertation, we study the existence and spatio-temporal symmetric patterns of peri-
odic solutions to second order reversible equivariant non-autonomous periodic systems with
multiple delays under the Hartman-Nagumo growth conditions. Our method is based on
the usage of the Brouwer D1 × Z2 × Γ-equivariant degree theory, where D1 is related to
the reversing symmetry, Z2 is related to the oddness of the right-hand-side and Γ reflects
the symmetric character of the coupling in the corresponding network. Abstract results are
supported by a concrete example with Γ = Dn – the dihedral group of order 2n