University of British Columbia
University of British Columbia: cIRcle - UBC's Information RepositoryNot a member yet
88081 research outputs found
Sort by
Seismic design and evaluation of self-centering modular rocking composite sandwich wall (SC-MRCSW) system
No full textWith rapid population growth worldwide and increasing urbanization, there is a growing demand for infrastructure that can be constructed quickly, while maintaining sustainability and resilience. In order to enhance the seismic resilience and constructability, this thesis introduces a novel self-centering modular rocking composite sandwich wall (SC-MRCSW) system. SC-MRCSW system consists of steel plates in-filled with concrete, where the steel plates are tied using steel bolts. The steel plates reduce the need for concrete formwork which significantly reduces the on-site constriction time, making SC-MRCSW system a highly efficient structural system for high rise application. To improve the performance of the SC-MRCSW, novel self-centering friction dampers are added to the system to ensure SC-MRCSW system can dissipate earthquake energy and self center after strong earthquake shaking. This makes SC-MRCSW system highly resilient. This thesis focuses on the seismic design of SC-MRCSW system. This is achieved using the novel Equivalent Energy Design Procedure (EEDP). Detailed numerical model of the SC-MRCSW system was developed to simulate the non-linear dynamic response of the system. The results show that SC-MRCSW system is able to achieve superior performance at different levels of earthquake shaking intensities. This thesis shows that the proposed SC-MRCSW system can be used as a reliable seismic force resisting system and can be designed efficiently using EEDP.Applied Science, Faculty ofCivil Engineering, Department ofGraduat
Language modeling techniques for biological sequence processing
No full textBiological sequences--DNA, RNA, and proteins--form the basis of genetic information in all living organisms and viruses, dictating everything from genetic inheritance to biochemical processes and physical characteristics. Accurate prediction and generation of these sequences are crucial for advancing personalized medicine, developing novel therapeutics, and understanding evolutionary processes. However, traditional language modeling techniques often struggle to achieve the level of precision required for real-world applications, where even a single bio-token error can render predictions biologically meaningless. This thesis addresses the central challenge of enhancing the precision of biological sequence processing through language modeling innovations that tackle different sources of prediction failure distinct from those arising in natural language generation tasks. We identify that biological sequence prediction suffers from both local semantic errors and global constraint violations. To address local errors, we introduce a reflection-based biological sequence pretraining framework that augments the autoregressive Transformer with self-correction capabilities. By incorporating auxiliary reasoning tokens and training the model to recognize and correct its own mistakes, we achieve significant improvements in amino acid precision and peptide-level accuracy. To address global constraint violations in biological sequence generation, we develop a non-autoregressive Transformer that leverages bidirectional global contextual constrained optimization. By incorporating a sequence-level precise mass control module, this approach achieves state-of-the-art results in protein sequencing tasks. These approaches demonstrate that biological sequence modeling demands domain-specific adaptations of language modeling techniques. Local error correction via reflection mechanisms addresses semantic and reasoning failures, while global constrained optimization enforces physical and chemical validity. Each method offers a distinct perspective on overcoming the limitations of naive natural language modeling in biological contexts.Science, Faculty ofComputer Science, Department ofGraduat
Prosociality "in the wild" : understanding prosocial engagement in everyday life
No full textPrior research suggests that prosocial behaviours (the diverse ways we act to benefit others) are generally associated with improved health and well-being outcomes for the person doing the helping. However, many studies show null findings, and some even find that prosocial behaviours are associated with adverse outcomes, potentially due to not accounting for other daily activities and other contextual variables. To address these gaps, I evaluated prosociality “in the wild”, or in the context of other prosocial behaviours (Study 1), other daily activities (Study 2), and the surrounding networks of prosocial persons (Study 3). In Study 1 (N = 40,817), I examined how diversity of prosocial engagement was associated with health and well-being. Prosocial engagement diversity was not associated with either life evaluation or self-rated health, and the total amount of prosocial engagement did not moderate these associations. There were differences across sociodemographic groups and when evaluating diversity within prosocial behaviour categories (e.g., diversity within types of care for non-household members). In Study 2 (N = 16,228), I examined the trade-offs in time use between prosocial behaviours and other daily activities using compositional data analysis. Trade-offs between prosocial and other activities were associated with health and well-being, especially depending on the activities involved, for whom trade-offs occurred, and other contextual factors (e.g., day of the week). In Study 3 (N = 119), I examined bidirectional associations between prosociality and social network integration in a 14-day intervention. Pre-intervention centrality was not associated with the proportion of acts of kindness performed during the intervention period, and performing a greater proportion of acts of kindness was not associated with post-intervention centrality. Network centrality was bidirectionally associated with the diversity of recipients of acts of kindness. Together, these findings illustrate that studying prosocial acts in combination with other activities, considering important sociodemographic moderators, and accounting for the recipients of prosocial acts, may influence associations between prosocial behaviours and health and well-being and present important avenues for future research. Further work is needed to inform and influence prosocial engagement recommendations that promote global prosociality in ways that enhance both individual and societal well-being.Arts, Faculty ofPsychology, Department ofGraduat
Les attitudes linguistiques et l'insécurité linguistique chez les locuteurs du français langue seconde à Vancouver (Language attitudes and linguistic insecurity amongst French second language speakers in Vancouver)
No full textCe mémoire porte sur les attitudes linguistiques des locuteurs de français langue seconde (L2) à l’Université de la Colombie-Britannique (UBC) et leurs sentiments d’insécurité linguistique. Cette étude permet de mieux comprendre comment les étudiants de français L2 à UBC acquièrent certaines attitudes envers le français et comment ces attitudes influencent leurs manifestations de l’insécurité linguistique. Pour aborder cette étude, j'ai mené des entretiens semi-structurés avec 11 participants. Les données recueillies lors des entretiens ont été analysées afin d'identifier les thèmes clés, offrant ainsi de nouveaux résultats sociolinguistiques sur les expériences des locuteurs de français L2 qui apprennent et parlent le français dans le contexte minoritaire français de la Colombie-Britannique. Les résultats révèlent que les attitudes linguistiques des participants sont façonnées par leur éducation dans les programmes d'enseignement du français langue seconde en Colombie-Britannique et au Canada. De plus, les attitudes linguistiques des participants, telles que la valorisation d'une variété standard du français de France, influencent leurs manifestations d'insécurité linguistique.Arts, Faculty ofFrench, Hispanic, and Italian Studies, Department ofGraduat
Embodying democracy in K–12 classrooms : wayfinding through rich and rough curricular terrain
No full textThis dissertation explores the possibilities of kindergarten to Grade 12 (K–12) education as a lived, democratic practice—dialogic, emergent, and grounded in co-constructed curriculum. Dewey’s understanding of democracy and education as a way of living together, more than just a governmental structure, offers a lens through which educators can interpret curriculum, pedagogy, and classroom communities. This research asks: What can we learn from educators who centre democratic habits and practices in their classrooms? What does it entail, and why does it matter?
Set in British Columbia, Canada, this research is situated within the 2016 K–12 curricular redesign, which marked a shift from content-driven, standardized education to a competency-based model emphasizing communication, collaboration, critical and creative thinking, and personal and social responsibility. While framed as a bold departure from instrumentalist approaches, its implementation has revealed tensions between aspirations and enduring practices. This research argues that years into this transformation, much can be learned from educators’ critical reflections on enacting its vision and their insights into its unrealized potential.
Guided by a qualitative, collaborative inquiry framework, this research engaged 14 K–12 educators over a school year in a series of dialogic gatherings. These educators, diverse in experience and context, participated in six in-person sessions and a public forum, collectively exploring democratic education through co-curricular wayfinding. Methodologically, the research draws on teachers as researchers, collaborative inquiry, narrative inquiry, currere, and embodiment, resisting the use of a single method in favour of a layered, process-oriented approach that foregrounds teacher agency and lived experience.
The conceptual metaphor of education as conversation is threaded throughout—from writing to research design to synthesis. Findings are mapped through the metaphor of democracy as rich and rough terrain, articulated across four interrelated features: Education as Conversation and Study, Reckoning With Imposed and Emergent Control, Cultivating Trust and Community—Nurturing Ecosystem Kinships, and Daring to Practise Freedom. Together, these illuminate the complex, dynamic spaces where democratic education takes root and evolves, offering insights into how educators and learners might navigate and nurture these landscapes together while revealing the transformative possibilities for learners and learning beyond standardized outcomes.Education, Faculty ofEducation, School of (Okanagan)Graduat
From blockchain trust models to native security protocols : enhancements of the A2A communication framework
No full textIn recent years, as multi-agent systems have rapidly expanded in open environments, trustworthy communication between agents has become a critical component of artificial intelligence infrastructure. Existing protocols face inherent trade-offs between security, interoperability, and decentralization: lightweight protocols rely on external trust assumptions and struggle to resist protocol-layer attacks; highly secure protocols feature complex structures that compromise practical usability. Constrained by this `agency communication trilemma', mainstream protocols generally lack native defenses against threats like registration spoofing, context tampering, and message replay under zero-trust conditions. Addressing this issue, this study uses Google's Agent-to-Agent (A2A) protocol as a case study to explore whether lightweight protocols can achieve native protocol-layer security while maintaining simplicity. The thesis first compares the security architectures of six representative protocols and identifies seven common protocol-layer threats across frameworks. Subsequently, it reverse-refines a minimal set of cryptographic primitives suitable for lightweight integration from threat analysis. These include signature verification based on decentralized identifiers (DIDs), nonce replay protection and context/historical integrity provided by chained and cumulative hashes. Although all of them are well-known and widely used in blockchain, this thesis is the first to show how they can be integrated into the A2A protocol without altering its original structure, embedding them directly into the message layer to strengthen communication security. Building on this enhanced design, the thesis develops a reproducible prototype and validates its effectiveness through simulated attack scenarios. Results demonstrate that the improved A2A effectively defends against major protocol-layer threats in zero-trust environments while preserving its original lightweight and interoperability advantages. This research provides a practical path for lightweight agent communication protocols that balance security and usability, laying the foundation for future trusted collaboration across cross-organizational, multi-domain agent ecosystems.Applied Science, Faculty ofEngineering, School of (Okanagan)Graduat
Health in the climate crossroads : exploring the integration of health in climate policy and practice in Canada
No full textThis doctoral thesis critically investigates the integration of health into climate adaptation policy and practice within the Canadian context. Through four empirical chapters, three of which have a particular focus on municipalities in British Columbia, it explores how health is understood, planned for, and acted upon in relation to climate change. This work is grounded in an analysis of interviews, planning documents, survey data, and a regional case study. Chapter 2 presents qualitative findings from interviews with local government leaders, examining their perceptions of climate-related health risks and their experiences with adaptation. It highlights governance gaps and calls for more multi-sectoral collaboration. Chapter 3 analyzes municipal climate action plans across British Columbia to assess how health is incorporated. While health risks are often acknowledged, the study finds limited integration of health systems or consideration of health co-benefits. The chapter recommends advancing a “Health in All Climate Policies” approach to strengthen cross-sector collaboration and public health responsiveness. In Chapter 4, survey data reveal that 23% of Canadians report experiencing physical health impacts from extreme weather events. Additionally, the survey shows that attitudes towards adaptation are influenced by a complex interplay of risk perception, health awareness, community belonging and demographic characteristics. The final chapter, Chapter 5, examines the implementation of cooling and cleaner air shelters in Metro Vancouver. Drawing on interviews with public and non-profit professionals, it identifies key lessons on service design, access, and the importance of sustained collaboration across sectors to protect vulnerable populations during extreme heat and wildfire smoke events. Overall, this thesis argues for stronger integration of health into climate policies and greater collaboration between health systems, local governments, and communities to protect public health in a rapidly changing climate.Science, Faculty ofResources, Environment and Sustainability (IRES), Institute forGraduat
Reliability analysis of deep learning accelerators via software-level fault injection
No full textDeep learning accelerators (DLAs) are increasingly deployed in
safety-critical applications such as autonomous vehicles and medical
diagnostics. However, these specialized chips are vulnerable to hardware
faults, including transient faults caused by radiation and permanent
faults due to aging or manufacturing defects. Existing fault injection
(FI) approaches face a trade-off between realism and scalability:
hardware-level methods, such as particle beam testing and
register-transfer-level (RTL) simulations, offer high accuracy but are
costly and slow, whereas software-level FI is efficient but often
inaccurate due to its lack of hardware awareness.
This thesis proposes a hardware-informed approach to software-level FI
that achieves both accuracy and scalability. We extract essential
hardware-level insights, such as realistic fault models and
microarchitectural characteristics, through a small, targeted set of
hardware-level studies. These insights enable the development of two
complementary software-level frameworks that realistically simulate the
effects of hardware faults in DLAs: (1) TPU-FI, which derives realistic
transient fault models from beam experiments on Google TPUs and
implements them in TensorFlow kernels, and (2) DLAFI, which uses RTL
simulations of systolic arrays to capture microarchitectural
characteristics and perform accurate permanent fault injection at the
LLVM IR level.Applied Science, Faculty ofElectrical and Computer Engineering, Department ofGraduat
Leveraging macrophage acidification for tuberculosis drug discovery
No full textTreatment for tuberculosis disease (TB) is failing on a global scale, necessitating new drug discovery efforts to identify direct and adjunctive therapies that reduce treatment complexity, length and side effects. Mycobacterium tuberculosis (Mtb), the obligate human pathogen and causative agent of TB, resides inside alveolar macrophages and combats acidic stress by residing in phagosomes and inhibiting their maturation. Mtb secretes a protein tyrosine phosphatase, PtpA, which binds to subunit H of the macrophage’s v-ATPase pump to block acidification. Improved understanding of the role of acidification in Mtb pathogenesis is needed and can assist in the development of better TB therapies. Following the hypothesis that acidification affects Mtb intracellular survival, I pursued the development of a phenotypic assay to identify compounds that modulate phagosome acidification in Mtb infected macrophages. The developed method uses Mtb intraphagosomal acidification to accurately correlate the proportion of live Mtb during infection. To evaluate the assay suitability for screening purposes, I screened two compound libraries and found that early acidification does not always correlate with inhibition of Mtb growth over time. Yet, early acidification occurs when directly interfering with microbial processes like DNA and protein synthesis over inhibition of cell wall synthesis. The link between phagosome acidification and Mtb viability was further studied using chemical inhibitors of v-ATPase function: Bafilomycin A1 (BafA1), Bafilomycin D (BafD), and Cladoniamide B (ClaB). I showed that all three limit the growth of intracellular mycobacteria in a bacteriostatic manner and that ClaB is metabolized by macrophages into a product with direct antimicrobial activity. I also found that the antimicrobial effect of BafA1 is abrogated in the Mtb ptpA mutant compared to parental strain and demonstrated PtpA:BafA1 binding. These findings suggest that although inhibition of phagosome acidification permits Mtb replication within macrophages, chemical inhibition may modulate host cellular processes leading to mycobacterial clearance. By assessing Mtb’s response to phagosome acidification and deacidification, this work sheds light on the complexity of host-pathogen interactions in early infection and provides a foundational assay that can be used to further investigate this phenomenon to uncover new direct and adjunctive treatments to eliminate TB.Science, Faculty ofMicrobiology and Immunology, Department ofGraduat
Roles of autophagy machinery components in the biology of Sclerotinia sclerotiorum
No full textSclerotinia sclerotiorum is a devastating necrotrophic fungus with a broad host range, infecting hundreds of plant species, including many economically important crops. Despite its significant agricultural impact, the molecular mechanisms that control its developmental processes and virulence are still not well defined. To better understand these mechanisms, a forward genetic screen was performed, which led to the identification of six UV-induced mutants harboring loss-of-function mutations in the autophagy-related genes SsATG1, SsATG2, SsATG4, SsATG5, SsATG9, and SsATG26. Functional validation through targeted gene deletions revealed that each of these ATG genes is essential for normal sclerotia formation. Although the overall morphology of appressoria remained largely unaffected, the mutants exhibited varying levels of impairment in vegetative growth and pathogenicity. Autophagic activity was further examined using the GFP–ATG8 reporter system, which demonstrated dynamic changes in autophagosome formation during development of sclerotia. The reduced or absent autophagic flux in mutant strains correlated with their inability to produce mature sclerotia and their attenuated virulence, emphasizing the importance of autophagy in fungal differentiation and infection. These results indicate that both macroautophagy and pexophagy play fundamental roles in nutrient recycling and energy allocation required for sclerotia maturation. Together, this study establishes a clear link between the autophagy machinery and the developmental and pathogenic processes of S. sclerotiorum. It provides a foundation for future investigations into specific cellular targets and regulatory pathways through which autophagy influences fungal virulence. Elucidating these mechanisms may ultimately contribute to the identification of novel molecular targets for controlling diseases caused by this highly destructive pathogen.Science, Faculty ofBotany, Department ofGraduat