- Thèses, 2011 - // Theses, 2011 -

Permanent URI for this collectionhttps://hdl.handle.net/10393/11105

Browse

Now showing 1 - 20 of 11136

The Influence of Exercise Training on Cardiovascular Health Outcomes in Adults with Cardiac Implantable Electronic Devices
(Université d'Ottawa / University of Ottawa, 2026-05-14) Roque Marçal, Isabela; Reed, Jennifer L.; Adamo, Kristi B.
Aims and Methods: The overall purpose of this dissertation was to examine the influence of exercise training on cardiovascular health outcomes in patients with cardiac implantable electronic devices (CIEDs). Study 1 retrospectively examined the changes in physical (e.g., cardiorespiratory fitness [CRF]), and mental health (e.g., anxiety levels) outcomes in patients with CIEDs completing an exercise-based cardiovascular rehabilitation (CR) program. Study 2 retrospectively assessed the associations of CR on major adverse cardiovascular events (MACE) among adults with CIEDs. Study 3 investigated, through a systematic review and meta-analysis, sex differences in CRF changes following exercise training in women and men with CIEDs. Study 4 tested the feasibility of a pilot randomized controlled trial comparing a 12-week virtual program of high-intensity interval training (HIIT) and moderate-to-vigorous intensity continuous training (MICT) in women with CIEDs. Results: Study 1 showed that patients with CIEDs (n = 252, 26% females) completing CR improved CRF and reduced anxiety and depression levels in patients with CIEDs. Study 2 found that CR was not associated with lower 5-year risk of MACE comparing propensity score-matched CIED patients with and without CR (n = 344, 23% females), whereas females derived greater benefit from CR. Study 3 demonstrated no sex differences in CRF following exercise among patients with CIEDs (n = 365, 22% women). Study 4 revealed that, among women with CIEDs (n = 20), a 12-week virtual HIIT intervention was feasible, whereas virtual MICT was not. Conclusions: Findings of this dissertation highlight (i) the importance of CR to improve CRF, a strong predictor of mortality, in individuals with CIEDs, (ii) the need of large-scale studies to understand the impact of CR on MACE in patients with CIEDs, (iii) a call to action to advance sex-specific inclusion and reporting practices in exercise trials with CIEDs, and (iv) remote HIIT is a feasible and safe exercise alternative to women with CIEDs, where results on key cardiovascular data can inform future trials.
Learning Interpretable Theories for Complex Neural Systems
(Université d'Ottawa | University of Ottawa, 2026-05-14) René, Alexandre; Longtin, André
Mathematical modelling has a long history of providing us with useful, succinct descriptions of the natural world. More than mere descriptions, the resulting models provide predictions that can be tested within the framework of the scientific method. But as we seek however to understand progressively more complex systems—especially in interdisciplinary fields like neurophysics—the methods of old don’t always yield the models we want: often they are too crude, or have too many free parameters—either of which makes them difficult to test. This work develops new paths to modelling, meeting complexity with today’s abundance of computational power. In contrast to other work however, which may outright replace the models of old with highly accurate but inscrutable ones, such as neural networks, here we stay firmly in the modelling tradition. By focussing on mechanistic, equation-based models, we extend rather than replace existing modelling capabilities. First we develop methods to learn complex models, where the form of the equations are known from our understanding of biology, but the effective parameters can only be inferred by fitting the entire model to data. These methods further allow to characterize correlations between parameters using Markov chain Monte Carlo (MCMC). However they can also lead to the proliferation of candidate models, which all plausibly fit the data. Therefore in a second part we develop a new approach to model comparison, integrating the principles of scientific induction with the methods of machine learning. The method is based on a novel random walk on the cumulative distribution function of errors; we thereby obtain an empirical modelling discrepancy (EMD) statistic that accounts for epistemic uncertainty on the model, specifically uncertainty under replications of the experiment. Together, these methods can be seen as the basis of a program for expert-driven, computer-enhanced research. And while developed in the context of computational neuroscience, they should be applicable wherever one is concerned with designing and studying mechanistic models, be they models in computational biology, statistical physics, or even econometrics. Les techniques de modélisation mathématique nous ont longtemps fournis des descriptions utiles et succinctes du monde naturel. Plus que de simple descriptions, les modèles ainsi obtenus fournissent des prédictions pouvant être validées à l’intérieur du cadre de la méthode scientifique. En cherchant toutefois à comprendre des systèmes de plus on plus complexes—en particulier dans les domaines interdisciplinaires comme la neurophysique—ces méthodes arrivent parfois à leurs limites: les modèles sont trop simplistes, ou possèdent trop de paramètres indéterminés—et par conséquent peuvent rarement être validés. Ce travail développe de nouveaux moyens de modélisation, répondant à la complexité avec l’abundance de puissance de calcul dont on bénéficie aujourd’hui. À l’inverse d’autres efforts toutefois, il n’est pas question ici de remplacer les modèles conventionnels par d’autres plus précis mais indéchiffrables tel que des réseaux neuronaux. En se concentrant sur des modèles mécanistiques, exprimés à l’aide d’équations, on reste plutôt fermement ancré dans la tradition de modélisation, étendant ses possibilités au lieu de les remplacer. On développe d’abord des méthodes pour apprendre des modèles, où la forme des équations découle de notre connaissance biologique, mais où les paramètres doivent être inférés en adaptant le modèle entier à des données observées. Une fois les paramètres appris, ces méthodes permettent ensuite également de caractériser leurs corrélations à l’aide de Monte-Carlo par chaînes de Markov (MCMC). Toutefois, elles peuvent aussi mener à un prolifération de modèles candidats, chacun ressemblant aux données. C’est pourquoi en seconde partie on développe une nouvelle perspective sur la sélection de modèles, intégrant les principes d’induction scientifique avec les méthodes d’apprentissage automatique. L’approache est basée sur un nouveau processus stochastique sur la distribution cumulative des erreurs; on obtient ainsi une statistique empirique de la divergence du modèle (qu’on nomme EMD, pour empirical modelling discrepancy). Cette statistique tient compte de l’incertitude épistemique du modèle, spécifiquement l’incertitude due aux réplications de l’expérience. Ensemble, ces méthodes peuvent être vues comme formant la base d’un programme de recherche mené par les experts et renforcé par le calcul. Et bien que qu’elles aient été développées dans le contexte de la neuroscience computationelle, elles pourraient être appliquées à tout contexte où l’on conçoit et étudie des modèles mécanistiques—comme la biologie computationelle, la physique statistique, ou même l’économétrie.
From Formal SYMBOLEO Specifications to Secure and Interactive Smart Contract Code
(Université d'Ottawa | University of Ottawa, 2026-05-14) Alfuhaid, Sofana; Amyot, Daniel; Mylopoulos, John; Anda, Amal
Context: Legal contracts have served as the bedrock of business transactions for millennia. They are core to modern supply chains, and their execution can now be automated through the use of i) smart contracts, supported by blockchain technology that safeguards data integrity, and ii) Internet-of-Things technologies to support their monitoring functions. Symboleo is a specification language used to formalize legal contracts, enable property analysis, and generate smart contracts for a permissioned blockchain platform (Hyperledger Fabric). However, automation around resulting smart contracts poses security challenges, particularly regarding who should have access to operate on contract elements. Additionally, how such smart contract should interact with their Cyber-Physical System (CPS) environment, including IoT devices, remains challenging. Purpose: The thesis proposes an architecture to integrate smart contracts, Complex Event Processing (CEP), message brokers, and a blockchain platform (namely Hyperledger Fabric) to support end-to-end Cyber-Physical Smart Contracts (CPSCs). This architecture makes it possible to connect IoT devices with smart contracts (generated using Symboleo) through a CEP engine and a message broker. Additionally, this thesis proposes an access control model, treating all contract elements as resources and ensuring regulated access by designated parties. This model extends the Symboleo ontology and language for legal contracts with new modeling concepts inspired by Role-Based Access Control (RBAC), tailored for the legal contract domain, resulting in SymboleoAC (Symboleo Access Control). SymboleoAC also extends the Symboleo language to handle dynamic contract execution scenario. Methodology: This research follows a Design Science Research methodology, which guides the development and evaluation of the research artifacts. This research is conducted in several iterative steps that are divided into two main phases, one that focuses on theoretical aspects and the other on the design, demonstration, and evaluation of the research artifacts. Contributions: The contributions of this thesis are: • An architectural framework for CPSCs that leverages complementary aspects of CPS and smart contracts; • SymboleoAC, an access control ontology for Symboleo; • An extension of the current Symboleo specification language (syntax and semantics) that supports smart contract requirements, including automation and control actions, access control, and CPS components; • An implementation of the SymboleoAC ontology and semantics into a reusable JavaScript library (SymboleoACJS), together with a tool, SymboleoAC2SC, that generates JavaScript smart contract code with security aspects for a designated platform (Hyperledger Fabric); and • A secure and event-driven SymboleoAC Application Programming Interface (API) that orchestrates the runtime ecosystem connecting IoT sensors, the message broker, the CEP engine, and the blockchain platform. Through extensive and the evaluation of multiple variations of two contract case studies, SymboleoAC (architecture, ontology, and language), along with its associated tools, is shown to be an effective environment for CPSCs, simplifying the design of secure smart contracts and their connections to message brokers, CEP engines, and IoT devices.
GOFinder-AI: Rapid and Explainable Gene Ontology Term Assignment Using Large Language Models
(Université d'Ottawa / University of Ottawa, 2026-05-13) Almir Ahmad, Aws; Mer, Arvind
Gene Ontology (GO) provides a structured vocabulary for describing the function of gene products. However, the rapid growth of biomedical literature makes manual GO curation increasingly difficult to sustain. Here, we present GOFinder-AI, a computational framework that supports literature-grounded GO annotation through pre-query text mining and large language model (LLM) inference. Given a biomedical text, the system identifies candidate GO annotations and produces supporting citations, explanatory reasoning, and linked biological entities. To improve task-specific performance, we fine-tuned multiple general-purpose LLMs (Llama-3.1-8B and Qwen3-8B) on a large, annotated dataset with more than 23,000 examples. Model performance was assessed using grouped 4-fold cross-validation, followed by evaluation on an independent test set containing >7000 gene-GO associations. Fine-tuning markedly improved performance compared to zero-shot prompting. The fine-tuned Qwen3-8B-based system reported higher predictive accuracy than GPT-5 mini, Llama-3.1-8B, and its own zero-shot counterpart. Overall, when tested on over 3,500 annotations, GOFinder-AI achieved a cumulative accuracy of 95.32%. It completed document-level GO curation in under one minute on average. GOFinder-AI offers a scalable, interpretable, and transparent approach to automated GO curation.
Context-Aware and Adaptive Multi-Scale Interest Modeling for Sequential Recommendation
(Université d'Ottawa / University of Ottawa, 2026-05-13) Wang, Xiaowen; Tran, Thomas
Sequential recommendation aims to predict a user's next interaction by modeling ordered user-item behavior sequences and plays a critical role in modern recommender systems. In real-world scenarios, user behavior is influenced by multiple contextual factors, among which temporal dynamics and item popularity are particularly important. Time intervals between interactions reflect the evolution and decay of user interests, while item popularity introduces frequency bias that may cause recommender systems to overemphasize popular items and underrepresent long-tail preferences. Moreover, user interests naturally exist at different temporal scales: long-term behaviors capture stable and persistent preferences, while recent interactions often reflect short-term, context-dependent intents. Effectively modeling these signals and integrating long-term and short-term interests remains a central challenge in sequential recommendation. To address these challenges, this thesis presents two sequential recommendation models with increasing modeling capability. The first model, Dual-Gated Time Frequency Co-Modeling for Sequential Recommendation (TiIfSRec), is designed to explicitly incorporate temporal intervals and item-frequency information into long-term and short-term interest modeling. TiIfSRec employs a dual-gated recurrent architecture in which time-interval signals control the decay of historical preferences, while item-frequency signals regulate the direction of state updates to mitigate popularity bias and preserve long-tail information. An attention mechanism is further introduced to highlight informative historical interactions and improve long-range dependency modeling. Experiments on multiple Amazon benchmark datasets demonstrate that TiIfSRec consistently outperforms representative time-aware and popularity-aware baselines, validating the effectiveness of jointly modeling temporal and frequency signals for sequential recommendation. While TiIfSRec improves long-term and short-term interest modeling, its fusion strategy relies on deterministic mechanisms, which limits flexibility when balancing stable preferences and rapidly changing intents. Motivated by this limitation, the second model, Diffusion-based Long-Short Interest Fusion for Sequential Recommendation (DiffLSRec), formulates long-short interest integration as a generative process. DiffLSRec introduces a diffusion-based framework in which the long-term interest representation serves as a generative prior, and the short-term interest representation is incorporated as conditional guidance during a multi-step denoising process. This progressive fusion strategy enables the model to adaptively adjust the contribution of long-term and short-term interests at different stages, rather than relying on a single static fusion operation. To further enhance contextual modeling and stability, DiffLSRec incorporates token-level contextual enhancement to capture fine-grained recent behavioral patterns, as well as a monotonic signal-to-noise ratio-adaptive guidance mechanism to regulate the influence of short-term signals throughout the diffusion trajectory. Extensive experiments show that DiffLSRec consistently outperforms strong sequential recommendation baselines, including diffusion-based models, across multiple evaluation metrics. Overall, this thesis demonstrates that explicitly modeling contextual behavioral signals and progressively integrating user interests across different temporal scales can substantially improve the accuracy, robustness, and adaptability of sequential recommender systems. The proposed models provide complementary contributions, with TiIfSRec offering effective context-aware interest encoding and DiffLSRec introducing a flexible generative paradigm for long-short interest fusion.
On 𝑝-adic 𝐿-functions Arising From Bianchi Modular Forms
(Université d'Ottawa / University of Ottawa, 2026-05-13) Deo, Mihir; Lei, Antonio
We study and construct 𝑝-adic 𝐿-functions of Bianchi modular forms, i.e., automorphic forms for GL₂ over quadratic imaginary fields, at non-ordinary primes in three different scenarios. In the first part, we construct signed two-variable 𝑝-adic 𝐿-functions with bounded coefficients from 𝑝-adic 𝐿-functions with unbounded coefficients for cuspidal Bianchi modular forms of parallel weight constructed by Williams. This construction extends the works of Pollack, Sprung, and Lei-Loeffler-Zerbes from the elliptic modular forms setting to the Bianchi modular forms setting. Additionally, we extend the results to 𝑝-adic 𝐿-functions coming from non-parallel weight 𝐶-cuspidal Bianchi modular forms constructed by Palacios. We construct logarithmic matrices using Wach modules basis constructed by Berger-Li-Zhu for the decomposition of 𝑝-adic 𝐿-functions with unbounded coefficients. We use Perrin-Riou's exponential map and the 𝑝-adic regulator to prove certain properties of logarithmic matrices. In the second part, we construct a 𝑝-adic Asai 𝐿-function, associated to a 𝑝-non-ordinary Bianchi modular form, which interpolates special complex 𝐿-values of the Asai 𝐿-function of that Bianchi modular form. This 𝑝-adic 𝐿-function has unbounded coefficients. We use modular symbols and some special cohomological elements, called Asai-Eisenstein elements, to construct polynomials. These polynomials satisfy some growth conditions, norm properties, and congruence relations. After taking the limit of these polynomials, we obtain the 𝑝-adic Asai 𝐿-function with unbounded coefficients. Moreover, we also construct signed 𝑝-adic Asai 𝐿-functions with bounded coefficients under some assumptions. In the third part, we construct a two-variable 𝑝-adic Asai 𝐿-function over the eigenvariety interpolating 𝑝-adic Asai 𝐿-functions of non-critical small-slope base-change Bianchi modular forms of parallel weight 0. To construct this 𝑝-adic 𝐿-function, we construct polynomials using a certain overconvergent modular symbol coming from a parallel eigenvariety associated with Bianchi modular forms and Asai-Eisenstein elements over an affinoid in a weight space. Their specialization at the weight (0,0) Bianchi modular form ℱ gives the 𝑝-adic Asai 𝐿 function associated to ℱ, which is constructed in the second part.
Exploring the Interaction Between Fibronectin and Transglutaminase II
(Université d'Ottawa / University of Ottawa, 2026-05-13) Grant, Thomas; Keillor, Jeffrey W.
Human transglutaminase 2 (TG2) is a versatile transamidating acyltransferase that is ubiquitously expressed in the human body. It catalyzes a host of functions ranging from crosslinking of glutamine and lysine residues via its transamidating activity in its calcium dependent "open" conformation to acting as a G-protein involved in signal transduction of G-protein Coupled Receptors (GPCRs) in its "closed" conformation. Another intriguing function of TG2 is its role as a cell surface protein where it interacts with a variety of extracellular matrix (ECM) proteins, forming a ternary complex between cell surface receptor β-integrins and the structural protein fibronectin (FN), an interaction important to FN deposition and fibrillogenesis in the ECM. Cell surface TG2 and its interaction with FN has been shown to play a role in tumour cell resistance to chemotherapeutics and increased adhesion of ovarian tumour cells to the ECM. Our objective was to contribute to the elucidation of the non-covalent interaction between TG2 and the 45 kDa Fibronectin gelatin binding domain (45FN) by analyzing potentially key residues on TG2 through alanine site directed mutagenesis (SDM) and Bio-Layer Interferometry (BLI). Additionally, we used Genetic Code Expansion (GCE) to incorporate the UV-photoactivable crosslinking unnatural amino acid (UAA) Azido-L-Phenylalanine (AzF) at these residues. We corroborated that R116 is an important residue for the interaction between 45FN and TG2, based on a three-fold increase in K_D when R116 was mutated to alanine, but we were unable to confirm results that support K30 as being crucial on its own. A crosslinked complex was formed between 45FN and TG2 with AzF incorporated at position F203 using the pULTRA-pCNF orthogonal tRNA/aaRS system. However, more optimization is required for this technique to be viable for analysis of the paired residues via crosslinking MS (XL-MS), due to the low concentrations of crosslinked complex that was formed.
Lateral Torsional Buckling of Built-Up Beams
(Université d'Ottawa / University of Ottawa, 2026-05-13) Mansor, Mohamed; Mohareb, Magdi; Doudak, Ghasan
Built-up timber beams are composed of multiple thin plies of equal depth, mechanically connected along their vertical interfaces using discrete fasteners to provide partial composite action. These beams are increasingly used in modern timber structures because thin plies are more economical and more readily available than wide solid sections. However, when such members are employed in long-span applications without adequate lateral bracing, their strength is often governed by lateral-torsional buckling (LTB). The LTB behaviour of these systems is further complicated by the presence of discrete mechanical fasteners, which provide only partial interaction between plies, rather than full composite action. Accurately quantifying this partial interaction and its influence on the critical moment remains a challenge in timber engineering and is not addressed in current American or European timber design standards. The former Canadian standard permitted the use of the LTB capacity of an equivalent monolithic section, provided that the plies were "securely connected." However, ambiguity surrounding what constitutes a secure connection prompted a shift to a more conservative approach in the most recent version of the standard, where the LTB capacity is taken as the sum of the individual capacities of the plies. Within this context, the present thesis investigates the elastic lateral-torsional buckling behaviour of built-up timber beams while accounting for partial interaction between plies. The first part of the study develops a three-dimensional finite element model to simulate the elastic LTB behaviour of two-ply built-up beams. The model captures relative slip between plies and idealizes fasteners as discrete springs at the ply interfaces by assigning shear stiffness in both longitudinal and transverse directions. This approach enables an investigation of the effect of fastener stiffness and spacing on the elastic LTB resistance and corresponding mode shapes of two-ply built-up beams. Building on this, the second part of the study introduces a variational principle and a beam finite element formulation for the LTB analysis of two-ply built-up timber beams. The formulation accounts for relative transverse and longitudinal slip between plies and characterizes fastener shear stiffness at their interface as linearly elastic springs, leading to an eigenvalue-type solution. The proposed solution yields results that are comparable to 3D finite element models but requires only a one-dimensional discretization along the member span, significantly reducing modeling, computational, and post-processing efforts. This finite element is then used for a parametric investigation of the effects of fastener stiffness, beam geometry, moment gradient, and load height on the elastic LTB capacity of two-ply built-up beams. The study also explores the potential of non-uniform fastener spacing, as opposed to conventional uniform spacing, as a means of optimizing beam capacity. The third part of the research extends the variational principle and finite element formulation to built-up timber beams composed of multiple plies. Predictions from the developed finite element model are verified against a 3D finite element model and full-scale experimental results reported by other researchers. The verified model is then used to conduct a comprehensive parametric study of 733 cases, investigating the effects of the number of plies, normalized fastener stiffness, longitudinal and transverse spacing, normalized span, ply aspect ratio, common load configurations, and load height on LTB resistance. The resulting database is used to develop dimensionless design equations through symbolic regression, characterizing elastic critical moments relative to the case of no interaction and deriving moment-gradient factors for common loading conditions and load-height coefficients. These proposed equations are integrated into a simplified design procedure, and their practical application is illustrated through design examples. In summary, the present study advances the understanding of LTB behaviour in built-up timber beams and provides practical tools for characterizing their elastic LTB resistance. The proposed solutions enable engineers to achieve more accurate and economical designs and lay the groundwork for future revisions of timber design standards regarding the LTB of built-up beams.
Regulation and Virulence Function of Fusarium graminearum Secondary Metabolites During Arabidopsis thaliana Infection
(Université d'Ottawa / University of Ottawa, 2026-05-12) Zouein, Marielle; Brauer, Elizabeth
During plant infection, Fusarium graminearum induces secondary metabolite production, which include mycotoxins and virulence factors, to overcome host defenses. Different hosts can elicit expression of distinct secondary metabolite biosynthetic genes through unknown host-derived signals. Here, we investigate how the Arabidopsis thaliana NADPH oxidase gene RbohD, which contributes to stress-induced reactive oxygen species (ROS) bursts, influences fungal secondary metabolite gene expression during seedling infection. F. graminearum infection induces progressive accumulation of ROS in Arabidopsis cotyledons which accompanied by tissue whitening. In the fungus, 650 genes are differentially expressed during infection including 51 secondary metabolite biosynthetic genes from the trichothecene, culmorin, fusaoctaxin, butenolide and fungal decalin-containing diterpenoid pyrone clusters (FDDP). In F. graminearum-infected rbohD knockout plants, cotyledon ROS accumulation was attenuated and fungal expression of genes from the trichothecene, fusaoctaxin, butenolide, and FDDP clusters were reduced. Knocking out other host susceptibility genes, including RLK7, ILK1, and APEX, had no effect on ROS accumulation and impacted expression of fewer fungal secondary metabolite genes. Abolishment of fungal production of deoxynivalenol, fusaoctaxin, and FDDP reduced F. graminearum virulence and resulted in increased callose production in the host during infection. Together, these findings indicate that the Arabidopsis RbohD gene influences F. graminearum pathogenesis and expression of secondary metabolite virulence factors during infection.
Intersectional Feminist Insights into the Lived Experiences of Domestic Violence and Trauma in Marginalized Women
(Université d'Ottawa | University of Ottawa, 2026-05-12) Sazgar, Raheleh; Lapierre, Simon
Violence against women continues to be a pervasive social issue worldwide, including in Canada, with women from marginalized social locations often facing compounded forms of violence, exclusion, and systemic neglect (Kaur & Garg, 2008; Singhal et al., 2021; Sokoloff & Dupont, 2005; World Health Organization, 2021). Domestic violence (DV) is a potentially traumatic experience for many women, often resulting in serious physical and psychological consequences, including post-traumatic stress disorder (PTSD) (Goodman & Epstein, 2008). However, the experiences and meanings that marginalized women attribute to trauma and PTSD remain underexplored (Baird, 2018). This thesis addresses this gap by critically examining how women who have experienced domestic violence and received a PTSD diagnosis understand and give meaning to trauma, and how they experience interventions provided by domestic violence and mental health services. Guided by an intersectional feminist framework (Crenshaw,1991), the research views trauma and violence as socially and structurally situated phenomena shaped by power, inequality, and identity (Brown, 2017; Herman, 2015). Using a feminist phenomenological approach and interpretative phenomenological analysis (IPA), the study draws on in-depth, open-ended interviews with women from diverse marginalized backgrounds in Canada, including Indigenous women and those from rural, religious, disabled, and sexually diverse communities. The methodological focus on lived experience and meaning making allows women’s voices to guide the interpretation of trauma beyond clinical or diagnostic frameworks. The findings highlight that participants conceptualize trauma not solely as an individual psychological response but as an ongoing experience of social and structural harm embedded within systems of inequality. Women described coercive control as a persistent form of domination that extends beyond physical violence and is shaped by race, class, disability, sexuality, and geography. Participants’ interpretations of PTSD reveal tensions between clinical definitions and lived realities of trauma, often exposing gaps in service delivery and cultural understanding. This research contributes to feminist scholarship on trauma and domestic violence by foregrounding the lived experience of marginalized women and challenging dominant, medicalized understandings of PTSD. It offers insights that can inform culturally responsive, intersectional, and trauma- and violence- informed practices within domestic violence and mental health systems.
Affordable Extended Hyperbolic Moment Closures for Rarefied Gas-Flow Predictions
(Université d'Ottawa | University of Ottawa, 2026-05-12) Rice, Ethan; McDonald, James G.
For the past 75 years, moment closures have been a promising method of gas-flow prediction for rarefied gases, as they offer significant mathematical and computational advantages over other applicable methods in regimes outside of local thermodynamic equilibrium. However, many of these advantages come from their ability to be formulated as hyperbolic systems of balance laws. Only recently have there been generalizable hyperbolic closures which can be expressed in closed form, and many of these closures have been restricted to simplified one-dimensional gases. While mathematically elegant, this limits the practical use of these new hierarchies to academic problems. Extending their desirable mathematical properties to real multidimensional gases has proven difficult, and a new method to handle this extension is the goal of this thesis. First, existing hierarchies of moment closures are presented, as well as their mathematical properties. Next, the technique by which higher-order moment models can be constructed is presented, with two new 20-moment closures being developed as a result. Linear stability of these models is presented, along with their performance in canonical discontinuous gas-flow problems for the continuum, transition, and free-molecular flow regimes. The traditional formulation of boundary conditions in kinetic theory is difficult to replicate in this framework. Instead, a new formulation of the Knudsen-layer boundary condition is presented, with results for both the 10-moment and 20-moment equations in canonical boundary-value problems. Finally, results for more realistic gas-flow problems in rarefied settings are shown. Strong shocks, and flows with regions of large translational non-equilibrium, are also explored. Further possible extensions for the models, such as for diatomic gases and plasmas, conclude the thesis.
Reaching Mobile Populations in Mass Drug Administration for Neglected Tropical Diseases: Evidence from Mali and Broader Implications for Africa
(Université d'Ottawa | University of Ottawa, 2026-05-12) Sangare, Moussa; Krentel, Alison; Yaya, Sanni
Neglected Tropical Diseases (NTDs) affect over a billion people globally, with mass drug administration (MDA) being one of the cornerstone strategies for control and elimination. However, mobile populations such as nomads, internally displaced persons (IDPs), migrants, and seasonal workers are frequently missed during MDA campaigns. This gap threatens progress toward the global targets for NTDs outlined in the WHO Roadmap 2021–2030, particularly in endemic regions like Mali and across Africa. Understanding the barriers to MDA participation among these groups is essential for designing equitable and effective interventions. Three complementary studies were conducted to explore MDA access among mobile populations. A cross-sectional study in Mali (2020–2021) used structured questionnaires and multivariable regression to identify factors associated with non-participation in schistosomiasis MDA. A qualitative study in Mali (2023) used in-depth interviews and focus group discussions to explore reasons for never being treated during MDA among mobile groups. Finally, a scoping review (2024/2025) followed PRISMA-ScR guidelines, synthesized evidence from 20 studies across Africa on mobility-related barriers to MDA. The following results present the main findings of our research. MDA coverage among mobile populations was consistently below the recommended 75% threshold. In Mali, only 40.8% of internally displaced people (IDPs) and 3.62% of migrants participated in the last MDA round. Key barriers included: lack of information (64.5%), geographic inaccessibility, mobility patterns (e.g., transhumance, seasonal work), low income and occupations such as mining, fear of side effects and rumors, and inflexible campaign schedules. Our findings also revealed that males, those facing physical or geographic barriers, and nomadic groups were significantly more likely to miss MDA. The scoping review highlighted additional systemic issues such as limited cross-border coordination and insufficient community engagement. Identified promising strategies included mobility-informed microplanning, flexible delivery models, and integrated health services. Mobile populations are systematically excluded from MDA programs, undermining efforts to eliminate NTDs. Addressing these disparities requires context-specific, adaptive, and participatory approaches. To ensure no one is left behind, MDA programs must move beyond one-size-fits-all models. Tailored strategies that account for mobility patterns, livelihood contexts, and local barriers are urgently needed.
"One Part Sane and Three Parts Mad": A Quantitative Study of Disability in Victorian Fiction
(Université d'Ottawa / University of Ottawa, 2026-05-11) Nash, Jessica; Gillingham, Lauren
This thesis uses computational techniques to examine how frequently Victorian authors rendered disability within their fiction. While past and present literary criticism of the Victorian novel has proven that disability plays a narratively important role in this era, my dissertation approaches this question computationally to analyze the relationship between disability and plot on a larger scale. In doing so, I argue that disability is not a modern-day concept applied anachronistically to the study of Victorian fiction, but a key feature of Victorian plots in an era where an expanding working class placed emphasis on the body and its ability to perform labour. To make this argument, I digitized hundreds of novels from the Victorian era and used randomized corpus downsampling to take a snapshot of the fiction published during this period. What the data reveals is that almost half of the novels contain the words "disabled" or "crippled" in them at least once. When I expanded my search-term list to include historically relevant words that ascribe disability using Henry Mayhew's London Labour and the London Poor (1851), the results show that close to 100% of the randomized fiction list included words related to disability. The frequency of disability in Victorian fiction highlights how the novel embodies shifting class ideologies through the legible forms of physical difference. This idea is proven effectively through the computational linguistic tools this thesis uses, such as natural language processing, which determined that the language of physical difference in novels is often found near mentions of poverty or class. What this co-occurrence demonstrates is that the boundary between poverty and disability was porous because of the tenuous relationship working-class Victorians had to income that was entirely dependent on their ability to perform and maintain labour. After making this argument, my thesis pivots from a distant reading to a close one, and I analyze select works of Charles Dickens and Wilkie Collins to argue that these authors use disability to furnish their narratives in contradictory ways that reveal their respective conceptions of the physical body in relation to its political and social world. My close readings use narratological concepts to study the structural affordances of disability within their fiction. By engaging with concepts from Disability Studies and crip narratology, I argue that Dickens flattens his disabled characters in service of plot momentum, while Collins shifts the weight of disability from identity to experience to create narrative contours that aid in the sensationalism of his novels. By studying the frequency of disability in over 3,000 works of Victorian fiction, in addition to the select works of authors known for their disability narratives, this dissertation reveals a key contradiction found at the heart of Victorian fiction: that the plots that work often rely on the bodies that can't.
Chinese Immigrant Women's Perspectives on Communication Strategies of Healthcare Providers During Gynecological Examinations
(Université d'Ottawa / University of Ottawa, 2026-05-11) Lu, Yutong; Cherba, Maria
Effective and appropriate communication is a key component of positive clinical experience, while immigrant women's experiences during intimate medical encounters remain underexplored. This thesis explores how Chinese immigrant women perceive, interpret, and engage in communication with healthcare providers during gynecological examinations in Canada. Drawing on Communication Accommodation Theory and patient‑centred communication, the study investigates how patients' emotional experiences, sense of agency, and perceptions of care quality are influenced by the communication strategies encountered. Using a qualitative research design, in‑depth semi‑structured interviews were conducted with twelve Chinese immigrant women who lived in Canada. Data were analyzed following Braun and Clarke's (2006) six-step thematic approach. The results were presented in three major themes: Communication Accommodation Strategies in Addressing Language Barriers; Patient-centred Communication in Alleviating Stress; and Cultural Beliefs Shaping Perceptions of Gynecological Care. Findings indicate that communication during gynecological examinations was co‑constructed through healthcare providers' strategies, patients' cultural backgrounds, and patients' active agency. Participants valued provider‑initiated accommodations and patient‑centred practices, such as linguistic adjustment, anticipatory explanations, privacy protection, and emotional reassurance, which helped reduce anxiety and foster trust. Importantly, patients were not passive recipients of care but actively prepared for and navigated examinations through information‑seeking, familiarisation with medical terminology, and in‑encounter questioning. Patient agency emerged as a significant theme shaping how communication strategies were interpreted and negotiated by participants. While cultural beliefs related to sexuality and hierarchy influenced expectations, their impact varied by length of residence, education, and health literacy, highlighting communication as a dynamic and relational process rather than a unidirectional provider‑led intervention. This thesis contributes to scholarship on healthcare communication by foregrounding immigrant women's voices in gynecological care and demonstrating how accommodation and patient‑centred practices intersect in intimate clinical contexts. The findings offer practical implications for improving communication training and promoting more humane, respectful, and empowering care for diverse patient populations.
Thickening Theory: Memoir, Affect, and Fat Feminist Futures
(Université d'Ottawa / University of Ottawa, 2026-05-11) Babin, Darby; Orsini, Michael
This dissertation is about the epistemological potential of fat life writing at the intersection of affect and embodiment. While the field of Fat Studies has its roots in liberation movements of the 1960s and 1970s (Wykes, 2014), contemporary 'body positivity' has commodified feminist ideologies to promote a sanitized fat politic that seeks participation in normative structures or what Da'Shaun L. Harrison (2021) calls "Desire/ability" (p. 12). Moreover, feminist studies has, at times, been epistemically ignorant around issues of 'fat', despite a feminist interest in the body and its meanings. As healthism and neoliberal ideals of individual responsibility have taken root, particularly through 'body positivity' and Health at Every Size (BP/HAES), a singular, "problematic model of fat subjectivity" (Murray, 2005, p. 155) has emerged. This project endeavours to contribute to ideas of messiness, ambiguity, and an acknowledgement of the contradictions of fat life, thereby challenging the assimilationism and achievement feminism (Farrell, 2021) at the centre of contemporary fat politics and thickening existing approaches to fat. Following the history of personal narrative in feminist theory, works by authors Roxane Gay, Lindy West, and Samantha Irby are analyzed in this thesis, highlighting the everyday affective encounters they archive with an emphasis on the role of race, queerness, and disability. Through their own stylistic and tonal choices, each author generates important contributions to fat studies' interest in cultural depictions of fatness, particularly the ways in which fat bodies are shaped by and shape the world. By mobilizing the work of Hil Malatino on "side affects" (2022), Sianne Ngai's "ugly feelings" (2005), Rosemarie Garland-Thomson's "misfitting" (2011), and Sara Ahmed's (2006; 2010) work on orientations and affect, this project aims to strengthen fat epistemology and contribute to a fattening of feminist studies which takes seriously the uncomfortable, the ugly, and the irreverent.
An Interpretable GeoAI Framework for Analyzing Multi-Ethnic Settlement Dynamics: Evidence from the Greater Toronto Area, 2001-2021
(Université d'Ottawa | University of Ottawa, 2026-05-08) Mashhadi Moghaddam, Seyed Navid; Cao, Huhua
This dissertation develops an interpretable Geospatial Artificial Intelligence (GeoAI) framework for understanding multi-ethnic settlement patterns in contemporary Canadian cities, demonstrating that computational sophistication need not sacrifice theoretical transparency or democratic accountability. Through three interconnected investigations spanning 52 major Canadian cities, 30,091 dissemination areas, and nine major ethnic groups (China, India, Philippines, Pakistan, Sri Lanka, Iran, Portugal, Italy, United Kingdom) over two decades (2001-2021), this research transforms multi-ethnic settlement from a descriptive sociological phenomenon into a predictive science grounded in interpretable physics-informed models. The research addresses a fundamental tension in urban artificial intelligence: while cities increasingly deploy algorithmic systems to manage complex urban dynamics, the dominant paradigm of black-box optimization systematically fails to achieve stated goals while potentially harming vulnerable communities. A critical analysis of 157 urban AI deployments (2015-2024) reveals pervasive "metrics traps" where impressive technical accuracy, such as ShotSpotter's 97% acoustic precision yielding only 9.1% crime-fighting effectiveness, consistently fails to translate into meaningful social outcomes. This critique establishes the ethical and methodological imperative for interpretable approaches in urban demographic analysis. The core methodological contribution is the development of a graph-based physics-informed neural network (GraphPDE) that embeds multi-ethnic reaction-diffusion dynamics while learning interpretable demographic parameters. Adapting Turing's pattern formation theory to spatial graphs, this framework reveals that ethnic settlement patterns emerge from self-organizing spatiotemporal processes with quantifiable characteristics: ethnic-specific spatial scales ranging from 34.5 km (Philippines) to 63.0 km (United Kingdom); pattern formation regimes segregating groups into spots (UK, Portugal), stripes (China, India, Philippines), and labyrinthine (Iran, Pakistan, Sri Lanka) morphologies; and critical nucleation thresholds varying from 658 individuals (China) to 8,132 individuals (India) for spatial clustering emergence. The learned attention-based interaction mechanism quantifies previously unmeasurable inter-ethnic dynamics, revealing that Chinese populations exhibit strong negative self-competition (-19.8) driving spatial dispersal while maintaining facilitative relationships with multiple groups, Philippines-Pakistan mutual attraction, and United Kingdom-Italy mutual repulsion. The Multi-Ethnic Spatial Mixture of Experts (MESMoE) framework synthesizes physics-informed modeling with machine learning, achieving state-of-the-art predictive performance (R² = 0.80-0.83) while maintaining complete interpretability through regime-specific expert modules for colonization, jump processes, decline, and continuous diffusion dynamics. This architecture demonstrates that incorporating domain knowledge enhances rather than compromises predictive accuracy, with physics-informed components accounting for 57.2% of prediction variance. The framework reveals systematic differences in how ethnic communities respond to urban infrastructure: Chinese and Filipino populations show amplification factors strongly correlated with transit accessibility (r = 0.58 and 0.61), while Indian populations demonstrate stronger correlation with housing variables (r = 0.47). Configuration landscape analysis identifies multiple stable settlement configurations with quantifiable transition barriers, revealing that demographic transitions require sustained interventions over decadal timescales due to asymmetric barriers creating lock-in effects. The temporal evolution of parameters captures non-stationary dynamics across four census periods, with Philippines-origin populations maintaining consistently positive growth rates (0.015-0.023 year⁻¹) while United Kingdom-origin populations transition from positive growth to sustained decline. This research makes four distinctive contributions: (1) developing progressive spatial analysis as a systematic methodology for studying complex urban phenomena through integrated multi-scale investigation; (2) demonstrating interpretable GeoAI that achieves competitive performance without sacrificing transparency; (3) revealing hidden ethnic dynamics through quantification of inter-ethnic interactions, nucleation thresholds, and pattern formation regimes; and (4) providing practical tools bridging academic research and urban planning practice through open-source implementations. The framework transforms abstract geographic concepts into measurable quantities, "sense of place" becomes configuration stability, "community cohesion" translates to interaction strength, and "neighborhood character" maps to position in pattern formation phase space. The policy implications are transformative: understanding asymmetric configuration barriers explains persistent settlement patterns while identifying intervention requirements; regime-specific approaches match policies to demographic dynamics; and cascade effects enable strategic investments generating system-wide benefits. Critical reflection acknowledges fundamental limitations, physics cannot capture individual agency, cultural meaning, or structural inequalities, yet the framework complements rather than replaces other ways of understanding urban dynamics. This dissertation demonstrates that interpretable GeoAI can bridge the persistent divide between quantitative spatial science and critical human geography, proving that mathematical rigor need not sacrifice social awareness. The discovery that Canadian cities' ethnic geography follows learnable physical dynamics while maintaining cultural distinctiveness suggests that diversity and order are complementary aspects of urban organization, offering both cautionary lessons about algorithmic governance and hopeful possibilities for creating more equitable, integrated multicultural cities.
Integration of New Eco-Friendly Armour Units into Coastal Structures
(Université d'Ottawa | University of Ottawa, 2026-05-08) Sayar, Serim Dogac; Nistor, Ioan
Growing demands for resilient and sustainable shoreline protection have led to the development of a new generation of armour systems that combine hydraulic efficiency with ecological performance. This thesis addresses this need by providing an empirical-based framework for the Coastalock armour unit, an eco-engineered alternative to traditional single-layer armour units. Through large-scale experiments, numerical modelling, and design guidelines, the study defines hydraulic performance and bio-enhancing design parameters that facilitate reliable, environmentally enhanced coastal defence solutions. The large-scale experiments examined the hydraulic performance of Coastalock armour units on low-crested and emergent rubble mound breakwaters under irregular wave conditions, quantifying wave transmission, overtopping, and breakwater stability. Numerical modelling was conducted using the IH2VOF CFD tool to reproduce some of the experimental conditions and investigate detailed hydrodynamic processes for the tested configurations. Sensitivity analyses of grid resolution, boundary conditions, and varying wave and structural configurations assisted the model’s accuracy and guided optimal parameter selection for Coastalock units in future simulations. The model simulated wave–structure interaction by providing free-surface elevations, wave transmission coefficients, and overtopping rates observed in laboratory tests, demonstrating its capability to capture complex wave–structure interactions. The final stage of the research consolidated 417 test runs from four multi-institutional experimental testing campaigns to develop the first comprehensive design recommendations for the Coastalock armour unit. Statistical and deterministic analyses were performed to quantify the effects of Coastalock-specific design parameters on armour stability. Armour unit spacing and underlayer size ratio emerged as the critical parameters for hydraulic stability. Design stability constants of Ns ≈ 2.8 and KD ≈ 15, determined for a 1V:1.5H slope, demonstrating that Coastalock performs comparably to well-established single-layer armour units. The design framework integrates biological performance by correlating ecological outcomes from ECOncrete’s monitoring of Coastalock installations with structural configuration variables, including unit orientation and spacing. Integrating these outcomes into the design process allows engineers to consider ecological functionality as a design objective alongside traditional hydraulic design. The research, therefore, establishes a data-driven foundation for designing dual-purpose coastal structures that satisfy both engineering and environmental targets, offering a practical pathway toward sustainable and resilient coastal infrastructure.
Learning Posted Prices in Bilateral Trade: Regret Guarantees Under Full and Bandit Feedback
(Université d'Ottawa | University of Ottawa, 2026-05-08) Bruni, Luca; Fraser, Maia
In this thesis we study an economically motivated sequential decision problem in which a learner repeatedly chooses an action (e.g., a posted price) and observes structured feedback. We ask how the information revealed after each decision determines whether learning is possible and what regret rates are achievable. We cast the problem in the online-learning framework and analyze two feedback models. Under full-feedback, the learner can effectively evaluate alternative actions; we give an efficient algorithm with sublinear regret and matching lower bounds, yielding sharp minimax rates. Under bandit- feedback, we show that without additional regularity, sublinear regret is impossible. We then identify natural smoothness conditions on the instance under which bandit learning becomes feasible again and derive regret guarantees. Overall, our results cleanly separate learnable from non-learnable regimes and quantify how mild structure can bridge the gap between full-feedback and bandit learning.
Suicide Ideation Detection from Social Media Using Language Models: Data Augmentation and Interpretability
(Université d'Ottawa | University of Ottawa, 2026-05-07) Ghanadian, Hamideh; Al Osman, Hussein; Nejadgholi, Isar
Early detection of suicide is a vital research area that holds great potential for facilitating early prevention and interventions by mental health professionals. With accurate and reliable detection of suicide ideation, targeted interventions can be developed to reduce suicide rates and provide better support for at-risk individuals. While traditional methods of identifying individuals at risk of suicide have primarily relied on clinical assessments and crisis hotlines, the ubiquity of social media platforms has opened new avenues for early detection and intervention, as many individuals at risk of suicide might express suicidal ideation in their social media interactions. However, developing models for suicide detection on social media is a challenging area of research, primarily due to ethical and practical issues in data collection and annotation. In this work, we investigate the potential and limitations of Large Language Models in addressing data quality and accessibility issues in suicide detection on social media. First, we explore the capabilities of the state-of-the-art generative LLMs as Zero-shot or Few-shot alternatives to classifiers trained with annotated datasets. Our evaluations of the ChatGPT system underscore the limitations of this model in detecting suicide notes and highlight the necessity of high-quality training datasets for fine-tuning specialized classifiers for this task. Then, we turn to assess the quality of existing datasets collected from social media. With this assessment, we seek to uncover the extent to which social media datasets mirror or diverge from conventional psychological understandings about suicide-related topics. We ground our evaluation of the datasets in established psychological literature by identifying risk factors linked to suicide, such as mental health challenges, relationship conflicts, and financial distress. Employing a guided topic modelling technique, we identify the distribution of mentions of risk factors in existing datasets. Our results demonstrate that while surface-level risk factors such as depression and anxiety dominate the topics of these datasets, more stigmatized topics such as racism, immigration challenges or sexual orientation prejudices are completely absent in these datasets. These results highlight the necessity of creating more diverse datasets that cover the risk factors related to under-represented social groups. Next, we focus on addressing the topic coverage issues in training datasets. Acknowledging that the sensitivity surrounding suicide-related data poses challenges in accessing diverse real-world examples, we introduce an innovative strategy that leverages the capabilities of generative AI models, such as GPT models, Flan-T5, and LLama2, to create synthetic data for suicidal ideation detection. Our data generation approach is grounded in social factors extracted from psychology literature and aims to ensure coverage of essential information related to suicidal ideation. Our comparison of synthetic and real data shows that synthetic data is more balanced in terms of risk factor coverage, is not significantly different from real data in terms of complexity and readability, and is significantly less diverse in terms of the vocabulary used. We then study the impact of psychology-grounded synthetic data on both the performance and the internal representations of suicide-detection models. We first leverage the generated synthetic data as standalone training data and as an augmentation source for fine-tuning BERT-family models for suicidal ideation detection. Our results show that synthetic datasets generated across multiple large language models enable strong generalization to real-world data, achieving an F-score of 82% when evaluated on held-out real samples. Moreover, augmenting this synthetic data with only 30% of the real dataset yields models that outperform those trained exclusively on the full real dataset, demonstrating a cost-effective strategy for improving performance while mitigating topic imbalance. Finally, we examine how topic-aware data augmentation influences the internal representations learned by these models. Using sparse autoencoders and geometric analyses, including UMAP projections and cosine-distance measurements, we analyze whether psychologically meaningful risk factors are encoded as more distinct and separable directions in the models' latent spaces. Our findings indicate that augmentation not only improves predictive performance but also leads to more structured and interpretable internal representations, with several previously under-represented risk factors becoming more clearly encoded. Together, these results highlight the value of combining synthetic data generation with representation-level analysis to develop more reliable and transparent models for suicidal ideation detection.
The Mediating Effect of Meaning Making on HEXACO Personality Factors and Internalizing Symptoms During COVID-19: A Longitudinal Analysis
(Université d'Ottawa | University of Ottawa, 2026-05-07) Holy, Celeste; Vaillancourt, Tracy
The COVID-19 pandemic was associated with increases in mental health difficulties, particularly for young adults. The literature remains unclear regarding whether differential coping strategies, as well as personality traits, influenced mental health outcomes over time. Thus, a cross-lagged panel analysis was used to examine the longitudinal effect of HEXACO traits and meaning making on internalizing symptoms across three time points during and after the COVID-19 pandemic. Results indicated strong autoregressive stability for both internalizing symptoms and meaning making across time. Cross-lagged analyses revealed that higher levels of internalizing symptoms at age 23 predicted lower levels of eXtraversion at age 24, while greater meaning making at age 23 predicted higher subsequent eXtraversion. There was no evidence that meaning making mediated the relationship between personality traits and internalizing symptoms over time. These findings suggest that meaning making may be more closely related to mental health outcomes within time points rather than functioning as a long-term mediating mechanism between personality traits and psychological distress. Moreover, elevated internalizing symptoms may constrain the expression of personality traits, while engaging in meaning-making processes may support the maintenance or expression of extraverted behavior over time. Clinical implications and directions for future research are suggested. Examining personality and meaning-focused coping within the context of a global crisis provides valuable insight into which individuals may be most vulnerable to psychological distress and which coping processes may promote adaptation in the face of acute and ongoing stressors.

Browse

Recent Submissions