Research Repository
LU|ZONE|UL distributes and preserves the scholarly work of LU faculty. It is a space for faculty to support the dissemination of knowledge created at Laurentian.
Electronic Theses and Dissertations (ETD) Repository This section preserves Master's theses and doctoral dissertations accepted at Laurentian University and is a mechanism for making this form of scholarly work widely accessible.
Communities in LU|ZONE|UL
Select a community to browse its collections.
Recent Submissions
Improving conservation translocation and ex situ breeding techniques for the recovery of a temperate zone rattlesnake
(Laurentian University Library & Archives, 2024-05-01) Choquette,, Jonahtan Daniel; Dr. Jackie Litzgus
Dr. Trevor Pitcher
Reintroduction biology is the study and practice of establishing populations of organisms using
tools like conservation translocation. An understanding of how to effectively conduct
translocations, particularly with snakes, is lacking and there is a need for technique
improvement. Systematic literature review, occupancy modeling, artificial hibernation,
retrospective analysis of zoo data, and population viability modeling were used to evaluate the
effectiveness of in situ and ex situ techniques to inform future translocations with temperate zone
snakes. Field work occurred at the Ojibway Prairie Complex and Greater Park Ecosystem in
southern Canada with the Eastern Massasauga (Sistrurus catenatus), a rattlesnake in decline
across its global range and in need of research into the effectiveness of population management
tools. When working with a low detectability species, results showed that evaluating long-term
success of translocations at achieving population establishment requires intensive survey efforts
to estimate patch occupancy across a study landscape. In the short-term, the invasive and sample-
size-limiting technique of radiotelemetry is required for evaluating success of release tactics. A
systematic literature review on snake translocations provided evidence for the utility of a suite of
translocation tactics for reducing postrelease effects. However, an assessment of ex situ
reproductive output coupled with demographic modelling, showed that certain beneficial tactics
(e.g., release of captive-reared snakes) can be problematic to implement and sustain, perhaps due
to breeding techniques that fail to replicate natural conditions, and with potential impacts to
release site fidelity of translocated snakes. Empirical and theoretical guidance were provided for
Massasauga recovery by informing eight steps for a successful snake translocation, beginning
with the establishment of translocation goals and ending with effectiveness monitoring.
Regardless, recovery efforts are hindered by the limitations of existing ex situ breeding
techniques, coupled with intensive supplementation efforts required to overcome postrelease
effects in situ and to establish a viable wild population over time. Research into alternative ex
situ breeding techniques for viperids (e.g., polygamous matings) and effectiveness of beneficial
translocation tactics for snakes in general are required. Reintroduction can be a daunting and
resource intensive pursuit, echoing the need to stabilize declining populations before they
become small.
Smart mining simulation with an intelligent supervisory agent for automated shovel dig allocation and truck deployment
(Laurentian University Library & Archives, 2023-11-20) Acheampong, Solomon Opoku; Dr. Eugene Ben-Awuah
The mining industry faces a substantial increase in data generation, making real-time
knowledge acquisition challenging. Mining data from day-to-day operations from
monitoring devices and sensors include the location of a truck or shovel, processing rate,
and quantity of material mined. The complexity of decision-making in mining has grown
due to intricate mineral deposits and the need for improved productivity.
The focus of this thesis is to develop a decision support framework referred to as smart
mining architecture (SMA), which leverages the power of simulation and reinforcement
learning for understanding mining system interactions and enhancing mining operational
performance. The mining operation and activities of mining equipment are simulated and
modelled using a discrete event simulation (DES) and an agent-based model (ABM)
respectively. The primary objective of this research is to develop an intelligent supervisory
agent capable of making smart decisions such as block sequencing, truck-shovel dispatching
and equipment maintenance and repairs towards improved operational performance. A
reinforcement learning deep quality network (DQN) algorithm is used to develop and train
the agent. The aim of the agent is to learn a policy that maximizes the expected return by
interacting in real-time with its environment taking smart actions to improve the learning
policy.
The developed intelligent supervisory agent (ISA) was successfully implemented for a
bauxite mine. The model was used to perform block sequencing, truck-shovel dispatching
and equipment maintenance and repairs in the SMA. Experimental cases with integrated ISA
were compared with experimental cases without the ISA. The output results with ISA
maintained consistent production levels in the presence of mining operational uncertainties.
Additionally, in the final experiment involving the DES, ABM and ISA, there was 3%
increase in total cashflow from the smart management of the short term mine plan and
mining equipment deployment as compared to experiments without the ISA. With the
introduction of the ISA, fewer workers are required to run the mining operation resulting in
financial savings. Fewer workers also mean less interaction between people and equipment
leading to improvement in health and safety records and occupational health challenges.
Enhancing neural mean teacher learning-based emotion-centric model for image captioning
(Laurentian University Library & Archives, 2023-11-09) Piramoon, Majid; Dr. Kalpdrum Passi
Image captioning is a task in computer vision and natural language processing that involves
generating a textual description of the content of an image. The goal of image captioning is to
create a system that can accurately recognize the objects, attributes, and relationships depicted
in an image, and generate a meaningful description of it in natural language, typically in the
form of a sentence or short paragraph. One of the state-of-the-art methods that we can use for
image captioning is Nemesis: Neural Mean Teacher Learning-based Emotion-centric Speaker.
Nemesis is a neural mean teacher learning-based emotion-centric speaker. It is a proposed
neural speaker capable of leveraging emotional supervision signals in the caption generation
process. Nemesis has been applied to the recently introduced ArtEmis dataset, which is the first
large-scale dataset for emotion-centric image captioning, containing 455K emotional
descriptions of 80K artworks from WikiArt. In this study, I employed a straightforward but
improved version of Self-Critical Sequence Training. By modifying the baseline function
choice in the REINFORCE algorithm, I introduced a simple alteration. The updated baseline
offers enhanced performance without any additional expenses, when compared to the baseline
that utilizes greedy decoding.
The role of preexisting anisotropies in focusing deformation in an Archean intrusion-related Au deposit: a case study from the Upper Beaver Au-Cu deposit, Ontario, Canada
(Laurentian University Library & Archives, 2023-09-09) Orlóci-Goodison, Ruth; Bruno Lafrance
The ca. 2680 Ma Upper Beaver deposit is an Archean intrusion-related gold-copper deposit located
in the southern Abitibi greenstone belt between Kirkland Lake and Larder Lake, Canada.
Mineralization is centered on the intermediate Upper Beaver Intrusive Complex which was
emplaced in the hanging wall of an extensional listric fault early during Timiskaming basin
formation. During subsequent deformation events, alteration, preexisting planar anisotropies, and
the orientation and composition of the ore zones enhanced strain partitioning, controlling the
development of folds, boudins and fabrics in strained ore zones. Steeply-dipping, sericite-altered
ore zones, oriented parallel to cross-stratal dikes, developed a continuous foliation surrounding
boudinaged and recrystallized quartz-calcite-anhydrite veins, whereas strong, shallowly-dipping,
stratiform, garnet-epidote-amphibole skarnoid ore zones developed a wavy, disjunctive cleavage
and deformed mainly by folding. The Upper Beaver deposit can be used as a guide for interpreting
the development of structures in similar but more complexly deformed deposits along major
deformation zones.
Diagnosis of pleural mesothelioma using machine learning
(Laurentian University Library & Archives, 2023-10-26) Abejide, Olaoluwa Julianah; Dr. Kalpdrum Passi
Mesothelioma is cancer that develops in the pleura. The most common cause of this disease is
contact with asbestos. Patients with mesothelioma have a better chance of surviving if they are
diagnosed quickly. This study utilizes a variety of machine learning to enhance pleural
mesothelioma diagnosis. The possibility of misclassification was decreased by extracting features
from a preexisting dataset. SVM, Decision Trees, and Random Forests are only a few machine
learning classifiers trained using essential and foundational features. Accuracy, precision, recall,
and F1-score were just a few measures used to evaluate these classifiers' performance in cross-
validation. SVM demonstrated excellent accuracy, precision, recall, and F1-score when classifying
individuals as either healthy or having mesothelioma. The results show the potential of machine
learning techniques for early diagnosis of pleural mesothelioma. Machine learning algorithms
improve diagnosis accuracy and turnaround time, improving patient outcomes. Using the results
of this research, a fully automated technique for diagnosing mesothelioma might be developed,
allowing clinicians more time to provide better care for their patients.