3789 research outputs found
Sort by
Towards a trustworthy data-driven clinical decision support system: breast cancer use-case
Get PDFArtificial Intelligence (AI) research has emerged as a powerful tool for health-related applications. With the increasing shortage of radiologists and oncologists around the world,
developing an end-to-end AI-based Clinical Decision Support (CDS) system for fatal disease
diagnosis and survivability prediction can have a significant impact on healthcare professionals as well as patients. Such a system uses machine learning algorithms to analyze
medical images and clinical data to detect cancer, estimate its survivability and aid in
treatment planning. We can break the CDS system down into three main components:
the Computer-Aided Diagnosis (CAD), the Computer-Aided Prognosis subsystem (CAP)
and the Computer-Aided Treatment Planning (CATP). The lack of trustworthiness of these
subsystems is still considered a challenge that needs to be addressed in order to increase
their adoption and usefulness in real-world applications. In this thesis, using the breast
cancer use case, we propose new methods and frameworks to address existing challenges
and research gaps in different components of the system to pave the way toward its usage
in clinical practice.
In cancer CAD systems, the first and most important step is to analyze medical images to identify potential tumors in a specific organ. In dense prediction problems like
mass segmentation, preserving the input image resolution plays a crucial role in achieving good performance. However, this resolution is often reduced in current Convolution
Neural Networks (CNN) that are commonly repurposed for this task. In Chapter 3, we
propose a double-dilated convolution module in order to preserve spatial resolution while
having a large receptive field. The proposed module is applied to the tumor segmentation
task in breast cancer mammograms as a proof-of-concept. To address the pixel-level class
imbalance problem in mammogram screenings, different loss functions (i.e., binary crossentropy, weighted cross-entropy, dice loss, and Tversky loss) are evaluated. We address
the lack of transparency in current medical image segmentation models by employing and
quantitatively evaluating different explainability methods (i.e., Grad-CAM, Occlusion Sensitivity, and Activation visualization) for the image segmentation task. Our experimental
analysis shows the effectiveness of the proposed model in increasing the similarity score and
decreasing the miss-detection rate. [...
Climate change effects on the surface temperature, ice coverage, & water levels of Lake Superior
Get PDFLake Superior’s surface water temperature, water level, and ice cover are all
suggested to be changing with increased atmospheric temperatures caused by climate
change. This thesis uses data from 1995-2022 to determine if similar results can be
found using a linear regression analysis. The only significant finds were of a yearly
water level increase and summer atmospheric temperature increase. However, it was
found that surface water temperature, ice cover, and atmospheric temperatures showed
similar patterns. More research is needed to further evaluate the impacts of climate
change on Lake Superior
Modulation scheme investigation for high-power medium-voltage current source converter based drives
Get PDFPulse width modulated (PWM) current source converter (CSC) based drives are commonly used in high-power (1-10 MW), medium-voltage (MV) (2.3-6.6 kV) applications.
These drives feature a simple converter structure, inherent four-quadrant operation capabilities, motor-friendly waveforms, and reliable fuseless short-circuit protection. PWM
CSC-based drives are generally constructed using symmetrical gate-commutated thyristors (SGCTs) with reverse voltage blocking capabilities. In order to avoid exceeding the
thermal limits of these SGCT devices, and to minimize switching losses, the device switching frequency used by PWM CSC-based drives is typically kept below 500 Hz. There are
three main modulation schemes used in MW-level MV PWM CSC-based drives: space
vector modulation (SVM), trapezoidal pulse width modulation (TPWM), and selective
harmonic elimination (SHE). Of these three modulation schemes, SHE possesses the best
harmonic performance as it features the ability to eliminate a number of low-order harmonics, all while retaining a low switching frequency. However, due to its off-line implementation, SHE suffers from poor dynamic performance, and in certain cases, requires a
large, memory-exhaustive look-up table. To address these issues, this research investigates
ways of improving the dynamic performance of conventional SHE through on-line (i.e.,
real-time) implementation. Two new modulation schemes are proposed: on-line SHE
for modulation of the grid-side PWM current source rectifier (CSR) and SHE-TPWM
for modulation of the motor-side PWM current source inverter (CSI). The proposed online SHE scheme models the independent switching angles used in conventional SHE as
polynomial functions by applying curve-fitting techniques. This method of implementation improves the dynamic performance of conventional SHE, as it enables real-time
computation of switching angles, and eliminates the need for look-up tables. Conversely,
the proposed SHE-TPWM scheme combines the principles, while retaining the respective
advantages, of both conventional SHE and TPWM. This integrative approach enables
SHE-TPWM to possess SHE-level harmonic performance, along with improved dynamic
performance rivaling that of TPWM
Monitoring and access management of resource roads with instrumentation
Get PDFForest transportation operations are facing challenges such as having larger road
networks to manage because of dispersed harvest patterns, climate changes forcing more
hauling to take place under wet road conditions, shortages of gravel, more worn-out roads
without a budget to rebuild, and the use of larger heavier trucks. This provides an
opportunity for the use of instrumentation with sensors and data acquisition systems to
allow real-time monitoring of road conditions that, when tied to threshold values, can be
used to manage access and control road operating costs. This research study will describe
the link between road material and road performance as well as different types of
instrumentation (how and why they are used). It will discuss a case study performed by
FPInnovations that used instrumentation to test if the installation of insulation within the
road structure of a weak and wet section of the road will improve road strength and
performance in the spring by preventing the road from freezing and thawing. Other uses
of instrumentation on resource roads will also be discussed as well as the use of
instrumentation in defining start and end dates for spring thaw load restrictions (SLRs) on
low-volume highways and resource roads across Canada
Malleable multiplication: the use of multiplication strategies and gamification to create conceptual understanding
Get PDFI have always considered math to be both my friend and foe. The parts of math I enjoy
are algebra, trigonometry, and multiplication. I could find solace in math that was formulaic
and had clear instructions. At times these concepts may be abstract, but there was always a
formula to consult. However, for every unit I found comfort in, there were always more that
distressed me. As much as math would stress me out and make me feel inadequate, I was
always trying methods to make it more enjoyable. I tried the placebo effect (which was not
successful because I knew it was a trick), positive affirmations, and bringing aesthetically
pleasing math accessories to class (i.e., pens, notebooks, etc.). Unfortunately, during my
elementary and high school years, I couldn't find a method that made learning math an
experience I would enjoy.
The moment everything clicked, and I found a method I could utilize came to me
during the second year of my Bachelor of Education degree. My teaching mathematics
professor tasked my class to invent or bring in a math game that would help students
understand a mathematical concept. I was interested in the project and had fun researching
and presenting my game. I was amazed by all the games my peers brought to class. Every
game was fun and helped instil concepts. I had never even considered gamification as a
method that could solve my stress and anxiety concerning math. I honestly didn’t even know
it was possible to teach such a wide array of mathematical subject matter using games as an
aid.
That game project sparked my interest in math and how the subject matter can be
made more accessible and fun for every student. After that, I started researching, reading
books about math education, and listening to podcasts. I got excited by this world where math
could be fun and inclusive instead of an anxiety-ridden subject. That one project has shaped
the course of my Master of Education program and inspired me to create a math game. I
would never have predicted that I would be focused on mathematics at any part of my
academic journey. For me, math has transformed from a scary subject to something
challenging but conquerable. I want every student to feel like they have the potential to
understand math and have fun with the subject. My fun with math may have happened
outside the confines of elementary and secondary school, but it still happened and has
changed a lot of my misgivings concerning the subject. [...
Computational efficiency maximization for UAV-assisted MEC network with energy harvesting in disaster scenarios
Get PDFWireless networks are expected to provide unlimited connectivity to an increasing number of heterogeneous devices. Future wireless networks (sixth-generation (6G)) will accomplish this in
three-dimensional (3D) space by combining terrestrial and aerial networks. However, effective
resource optimization and standardization in future wireless networks are challenging because of
massive resource-constrained devices, diverse quality-of-service (QoS) requirements, and a high
density of heterogeneous devices. Recently, unmanned aerial vehicle (UAV)-assisted mobile edge
computing (MEC) networks are considered a potential candidate to provide effective and efficient
solutions for disaster management in terms of disaster monitoring, forecasting, in-time response,
and situation awareness. However, the limited size of end-user devices comes with the limitation
of battery lives and computational capacities. Therefore, offloading, energy consumption and computational efficiency are significant challenges for uninterrupted communication in UAV-assisted
MEC networks. In this thesis, we consider a UAV-assisted MEC network with energy harvesting (EH). To achieve this, we mathematically formulate a mixed integer non-linear programming
problem to maximize the computational efficiency of UAV-assisted MEC networks with EH under
disaster situations. A power splitting architecture splits the source power for communication and
EH. We jointly optimize user association, the transmission power of UE, task offloading time, and
UAV’s optimal location. To solve this optimization problem, we divide it into three stages. In the
first stage, we adopt k-means clustering to determine the optimal locations of the UAVs. In the
second stage, we determine user association. In the third stage, we determine the optimal power of
UE and offloading time using the optimal UAV location from the first stage and the user association
indicator from the second stage, followed by linearization and the use of interior-point method to
solve the resulting linear optimization problem. Simulation results for offloading, no-offloading,
offloading with EH, and no-offloading no-EH scenarios are presented with a varying number of
UAVs and UEs. The results show the proposed EH solution’s effectiveness in offloading scenarios compared to no-offloading scenarios in terms of computational efficiency, bits computed, and
energy consumptio
User-centric clustering and pilot assignment in cell-free networks : a stochastic optimization approach
Get PDFCurrent 5G networks, primarily built on the cellular massive MIMO physical
layer technology, achieved significant improvement in spectral efficiency as compared to previous generations. Nevertheless, there is always an increasing demand
for higher data rates, and more reliable and uniform service. After successful massive MIMO deployments, it has become a natural question, "what will the physical
layer in beyond 5G and 6G networks be like?"
Cell-free massive MIMO has emerged as a promising physical layer technology
for supporting future deployments in beyond 5G and 6G networks. The main concept is to go beyond the cellular paradigm by employing an ultra dense deployment
of small-sized multi-antenna access points (APs) which cooperate to serve users in
the coverage area, eliminating the notion of boundaries between cells. The cell-free
architecture has shown the capability of providing uniform service within the coverage area, while cellular networks suffer from poor performance at cell edges. It also
has better ability to manage interference due to cooperation between APs which is
not the case in cellular networks with no cooperation.
The most practical form of this paradigm is user-centric cell-free massive MIMO.
Instead of allowing all the APs to serve all the users in the network, each user is
served by a subset of the APs which ensures that network operation is scalable as the
number of users grows. The main objective of this thesis is to provide a structured
approach to design the cluster of APs that serve each user which is known as the
user-centric clustering problem. On the pursuit to solve the clustering problem, there
is another problem which is tightly connected to it, the pilot assignment problem. Both
problems must be solved together to ensure satisfactory network-wide performance. [...
Memory, perception, and evaluation of emotional stimuli: the effects of oral contraceptive use, sex, and gender
Get PDFLittle research has examined if oral contraceptive (OC) mood side effects might be due to OCrelated effects on affective judgements or memory for emotional stimuli. Previous studies on sex
differences in emotional processing have rarely examined continuous gender (e.g., masculinity)
or OC-related sources of variation. In this lab-based study, OC users, free-cycling women (i.e.,
nonusers), and men rated the emotional valence and intensity of emotional stimuli across three
sensory modalities (e.g., visual, auditory, olfactory) to assess their immediate perception,
evaluation, and memory for the stimuli. Differences in ratings were examined as a function of
sex, masculinity, and OC use. In terms of emotional memory, OC users recalled more positive
and less negative information than nonusers (i.e., relatively more positive than negative words,
fewer negative objects and negative words). In terms of valence ratings, OC users and nonusers
differed in their overall perception of stimuli, but the direction was stimulus-specific. Compared
to non-users, OC users were more likely to perceive odours as positive and words as negative,
and more likely to perceive negative facial expressions and negative words as negative. In terms
of affective intensity ratings, OC users evaluated stimuli overall as more intense than nonusers,
with this group effect being driven by olfactory intensity ratings. There was no evidence that
gender (i.e., self-reported masculinity or measured voice pitch) explained a significant amount of
variance in women’s affective valence or intensity ratings of stimuli, although women’s voice
pitch was positively correlated with their olfactory intensity ratings. The OC-related emotional
memory effect, stimulus-specific valence bias, and enhanced affective intensity bias are
discussed in relation to findings from previous studies examining hormonal factors in emotional
processing
Effects of gratitude and cognitive load on delay discounting: replication failures in two experiments
Get PDFDelay discounting is the phenomenon whereby the value of future rewards is discounted as a
function of time. Individual differences in discounting rate have been linked to a range of
correlates and research has suggested a lower discounting rate to be more adaptive. One
mechanism that may reduce discounting rate involves effortful self-regulation achieved through
the engagement of executive function processes. This mechanism, however, is reliant on a
limited-capacity cognitive system. Cognitively demanding contexts and low baseline capacity
therefore create vulnerability to higher discounting rates and the associated negative sequelae.
The affective state of gratitude has been proposed as an alternative mechanism to reduce
discounting rate. It has been described as independent of effortful self-regulation with the
implication that it is not demanding of limited cognitive resources. However, this had not been
tested experimentally. The current research program comprised two experiments. The primary
aims were as follows: (1) to replicate previous findings showing the effects of gratitude and
cognitive load on discounting rate, and (2) to extend previous findings by investigating whether
the effect of one of these predictors depends on the level of the other. [...
Emerging adults’ perceptions of their family systems: resilience and connections after the COVID-19 pandemic
Get PDFFamily systems can be conceptualized as complex adaptive systems consisting of
intricate interconnections among family members that adapt dynamically to the environment.
The COVID-19 pandemic was a chronic and persistent trauma to many systems including
families. Family resilience, an ongoing process of the system that helps the family adapt to
changes and find an improved level of functioning, may have contributed to how families are
faring after the COVID-19 lockdown period. Purpose: This study focused on family resilience
as a possible moderator between COVID-19-related changes and family satisfaction, and how
structural dynamics of family connections can affect family resilience. Methods: N = 149
emerging adults ages 18-29 in Thunder Bay, Ontario, completed an online survey regarding their
family of origin, responding as a child within the family system. Various scales assessed the
number of COVID-19 stressors, COVID-19 impacts on the family, family resilience, family
satisfaction, and the frequency of connections with each family member. [...