39339 research outputs found
Sort by
Temporal Pyramid Structure for Video Frame Interpolation
No full textThe most prevalent structure in video frame interpolation involves using optical flow to guide frame warping, which typically considers only the two adjacent frames. However, these methods often fail to capture long-range temporal dependencies and often result in significant deformation in complex motion scenarios. We propose a novel Temporal Pyramid Attention (TPA) block, which employs a temporal pyramid structure to connect four frames within a sliding window for the generation of intermediate frames. The temporal pyramid structure consists of three layers to leverage multi-level features, estimate the frame window, and connect with a GRU to generate a bi-directional feature flow. Furthermore, the dual pyramid structure incorporates channel attention mechanisms, enabling the interpolation of three frames in a single process. The TPA block employs a multi-scale approach to effectively capture temporal dependencies and spatial correlations, enhancing the quality of interpolated frames. Our model achieves a state-of-the-art performance on the Vimeo90K septuplet dataset compared to existing methods using pre-trained parameters
Interpretation and classification of FBN1 variants associated with Marfan syndrome: consensus recommendations from the Clinical Genome Resource’s FBN1 variant curation expert panel
No full textAbstract Background In 2015, the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) developed standardized variant curation guidelines for Mendelian disorders. Although these guidelines have been widely adopted, they are not gene- or disease-specific. To mitigate classification discrepancies, the Clinical Genome Resource FBN1 variant curation expert panel (VCEP) was established in 2018 to develop adaptations to the ACMG/AMP criteria for FBN1 in association with Marfan syndrome. Methods The specific recommendations were developed through literature review, surveys, online expert panel discussions, and pilot testing of a set of 60 different variants. Consensus among experts was considered reached if at least 75% of the members agreed with a given rule specification. The final set of rules received approval from the ClinGen Sequence Variant Interpretation Working Group. Results The developed specifications introduce modifications to 14 of the 28 ACMG/AMP evidence criteria and deem 6 criteria non-applicable. Some of these specifications include refining the minor allele frequency thresholds, creating a FBN1-specific flowchart for PVS1, defining functional domains of the protein, developing a point-based system of counting probands and instances of de novo occurrences, recommending a points-based method of accounting for family segregation data, and clarifying the applicable functional assays that should be considered. To date, this VCEP has curated 120 variants which have been deposited to ClinVar with the 3-star review status. Conclusions Establishing specific adaptations for FBN1 has provided a framework to foster greater classification concordance among clinical laboratories, ultimately improving clinical care for patients with Marfan syndrome
Leadership Behaviours and the Perception of Support for Wellness Among Front-Line Healthcare Providers in Long-Term Care: A Post-Pandemic Recovery Effort
No full textThe COVID-19 pandemic brought to public attention extensive deficiencies in the long-term care (LTC) sector. LTC in Canada has been struggling with several challenges that reduce the quality of care delivered to residents, such as a lack of funding and standardized regulations. Front-line healthcare providers working in LTC settings have had to deal with work overload and drastic staffing shortages while trying to maintain their own overall well-being. While these workers are facing elevated stress levels, increased burnout symptoms, emotional exhaustion, and daily risks to their physical safety, it is the responsibility of direct managers and supervisors to help support them to the best of their abilities. The perceived availability of social support is associated with workers' positive well-being and leadership behaviours can impact workers' overall health. Thus, I used the social support and buffering hypothesis theory to explore what leadership behaviours can have an impact on perceived social support (PSS) among LTC workers. I used an intrinsic and instrumental case study design, which included reviewing documentary evidence and conducting in-depth, semi-structured interviews with nurses and personal support workers. The study setting was the Perley & Rideau Veterans Health Centre in Ottawa, Ontario. Collected data was analyzed using reflexive thematic analysis to answer two research questions: (1) What leadership behaviours do front-line healthcare providers perceive as being supportive of their overall well-being? (2) What types of social support (emotional, informational, social companionship, and instrumental) do front-line healthcare providers perceive as important or wish to receive from their leaders? Reflexive thematic analysis revealed four main overarching themes for supportive leadership behaviours: (1) relationship-building behaviours; (2) trust-producing behaviours; (3) safety-promoting behaviours; and (4) health-promoting behaviours. Health-promoting behaviours were most prominent in the data, likely because they are not isolated actions but are deeply intertwined with how leaders build and maintain relationships, establish trust, and create a safe working environment. Across all overarching themes, emotional support was particularly poignant to employees. This suggests that emotional support may be foundational to employees' overall perceptions of supportive leadership. The thesis results provide insights that can enhance our understanding of leadership behaviours and PSS in LTC and offer valuable recommendations that can be applied in LTC settings
Exploring Knowledge and Perspectives of Artificial Womb Technology in Canada and Australia: A Qualitative Multi-Methods Study
No full textA century ago, J.B.S. Haldane brought to life the concept of assisted reproduction by envisioning ectogenesis, that is gestation outside of the human body, as the future of science. Since then, the concept of artificial womb technology has featured in bioethical debates as well as in the public imagination through popular culture. Perhaps most importantly, the potential of artificial womb technology has influenced the development of related public policy all over the world. More specifically, in both Canada and Australia, the Assisted Human Reproduction Act and the Research Involving Human Embryos Act, respectively, criminalize ectogenesis. Recent advances in the development of precursors to artificial womb technology and related technologies have resulted in renewed interest in academic communities with respect to the potential implications of this technology. Nonetheless, much of the substantive debate and discussion on artificial womb technology over the last decade has happened beyond the reach of the general public. Decision-makers in Canada and Australia need to be equipped with data that shed light on views of the citizenry in the event that efforts to modernize legislative frameworks move forward to reflect current technological, health, legal, and socio-ethical realities. This thesis uses qualitative methods to explore the knowledge and perspectives of Canadian and Australian citizens with respect to artificial womb technology. In addition, this thesis examines the potential implications of artificial womb technology for the 14-day rule which prohibits embryo transfer, the impact on highly stigmatized health services, such as abortion, and already marginalized sexual and gender diverse populations
A Fragmented System: Understanding Child Protection Decisions for Canadian Parents with Intellectual Disabilities
No full textParents with intellectual disabilities are disproportionately represented in child protection investigations in Canada and around the world. In some cases, children are removed from mothers with intellectual disabilities soon after birth. Research has shown that parents’ intellectual disability status often plays a role in child protection decisions. However, there is limited research on what specific conditions or factors impact the final decision to remove child custody or reunify a family headed by parents with intellectual disabilities. In an effort to comprehensively examine the conditions that influence child protection decisions involving Canadian parents with intellectual disabilities, this dissertation undertook three qualitative studies: 1) content analysis of published court cases in Ontario; 2) semi-structured interviews with child protection workers from local agencies in Ontario; and 3) semi-structured interviews with parents with intellectual disabilities in Ontario and Manitoba. The first study examined how child protection decisions were made at family court and how the COVID-19 pandemic impacted child protection decision-making for parents with intellectual disabilities in 10 published Ontario court cases. The findings of this study contributed to the extant literature showing judicial reliance on parents’ intellectual disability to permanently remove child custody. Furthermore, this study highlighted cognitive accessibility challenges and inconsistent use of “evidence-based” practices in the child protection system. The second study captured the perspectives of 11 child protection workers supporting parents with intellectual disabilities in Ontario through semi-structured interviews. The findings of this study revealed the immense need for disability-specific trainings for child protection workers and the need for collaboration across developmental service agencies and child protection agencies. This study also had a promising finding regarding child protection workers not relying solely on parents’ intellectual disability status in their decision-making and taking intersecting factors into consideration. The third study used an inclusive research design and included 13 parents with intellectual disabilities in Ontario and Manitoba. Their perspectives were analyzed in collaboration with a co-researcher with lived experience. The findings of this study showed the importance of informal social supports and the need for continuity of formal social supports. Some parents named the child protection agency itself as an instrumental support and reported having positive relationships with their workers, in contrast with some previous research. The overall findings of the three studies are discussed using the Inequities Contextual Framework by Powell et al. (2024) to demonstrate that barriers at the contextual and institutional levels heavily influence outcomes for families headed by parents with intellectual disabilities. The implications for service system changes are also discussed
A quality improvement initiative aimed at reducing service strain and improving physician wellness in internal medicine
No full textAbstract Introduction Hospital strain has been shown to negatively impact physician wellness, educational experience, and patient care. To address rising service demands, a non-academic hospitalist service was implemented to reduce daily clinical teaching unit (CTU) census by approximately 30%. Secondary aims were to evaluate physician and trainee wellness on CTU as well as assess unintended adverse patient outcomes. Methods A two-phase intervention was implemented at one of two academic hospital campuses in January and April 2023. Mean daily census, mortality, 30-day readmissions, and length of stay (LOS) were obtained from an administrative database for the pre-study (October to December 2022) and study (January to December 2023) periods. The Mini-Z physician wellness survey was administered in March, June and December 2023. Data were analyzed by quarters using descriptive statistics as well as parametric and non-parametric testing, and a reflexive thematic analysis was undertaken. Results A CTU census trough of 71.3 was briefly attained in the second quarter of 2023 but increased to 78.6 in the fourth quarter of 2023, while remaining below pre-intervention levels. The proportion of attendings and residents reporting burnout was significantly different at the intervention (65.2%, n = 15/23) versus non-intervention site (94.1%, n = 16/17) in Q4 2023 (p = 0.033). Burnout was positively correlated with daily CTU census across both sites (r = 0.906). There were no differences in proportion of in-hospital mortality (p = 0.854), 30-day readmissions (p = 0.262), or LOS (p = 0.977) between the pre- and post-implementation periods. Qualitative analysis identified the hospitalist program as beneficial, but inadequate to address workload, education challenges, and patient safety concerns. Conclusion The addition of a non-academic hospitalist service reduced CTU census numbers and improved burnout, but the improvement in service strain was limited by rising admissions. Multifaceted approaches to wellness are needed, but this study supports ongoing endeavors aimed at reducing clinical workload to optimize the clinical teaching environment
Translating for Palestine as Self-development: A Translational Perspective on Human Rights and International Development
No full textThis thesis examines the translational practices of Local Non-Governmental Organizations (LNGOs) in Palestine and their role in implementing human rights development projects. By conducting semi-structured interviews with representatives from various LNGOs and engaging in an interdisciplinary dialogue between Translation Studies and Development Studies, the research analyzes how these organizations translate and adapt international development concepts to align with local needs and values. This study critiques the conventional perception of translation as a solely linguistic transfer, which often presumes the universality of human rights concepts and overlooks the complexities of cultural and contextual differences. It argues for an expanded definition of translation that encompasses deeper engagement, negotiation, and reciprocity in intercultural exchanges.
This research positions translation as a critical epistemic paradigm for understanding and managing the flow of knowledge, values, and terms of references between different contexts, particularly in asymmetrical power dynamics characteristic of international development. By reconceptualizing translation as both an archetype or a perspective and an analytical tool, this study explores how LNGOs navigate tensions between donor expectations and local realities, often reframing human rights principles to resonate more effectively with the Palestinian context. This process extends beyond linguistic conversion to include broader discursive and strategic engagements, highlighting the agency of LNGOs in shaping development discourse.
The findings demonstrate that LNGOs’ translational practices contribute to a more nuanced understanding of human rights and development, emphasizing the potential for translation as a broader concept to facilitate self-development and empowerment. This thesis explores an interdisciplinary approach that builds on the common assumptions and paradigms underlying both TS and DS to foster a participatory, context-sensitive development model. By centering local voices and perspectives, this model aims to support Palestinians' self-determination and enhance the effectiveness of human rights initiatives
Heuristic Approaches for the Home Health Care Staff Scheduling Problem
No full textThe increasing demand for community-based healthcare services requires efficient scheduling of caregivers to ensure high-quality, consistent care for clients while minimizing costs. This thesis addresses the challenge of creating optimal schedules for Personal Support Workers (PSWs) in supportive housing complexes, focusing on maximizing consistency for clients and reducing reliance on external staffing agencies. The problem is formulated as a Mixed Integer Linear Programming (MILP) model, incorporating specific objectives and constraints that assign PSWs to patient requests. Exact methods such as Branch and Bound (B&B) and Branch and Cut (B&C) were initially explored but, due to the complexity of the problem, these proved incapable of providing timely solutions. To overcome this, heuristic approaches, including Logic-Based Benders Decomposition and Lagrangian Relaxation were employed, where the solution process was stopped prematurely, effectively truncating the solution. As a case study, the proposed model is applied to Nucleus Independent Living (NIL), an organization based in Oakville, which aims to improve its scheduling practices for at-home and supportive housing services. The computational results demonstrate that Benders Decomposition outperforms Lagrangian Relaxation, achieving a solution within 0.2% of the best result obtained by the commercial solver GUROBI
Advancing Cross-Domain Fake News Detection: Enhanced Models to Improve Generalization and Tackle the Class Imbalance Problem
No full textThe rapid proliferation of fake news across domains, such as politics, health, and social media, poses a significant threat to the integrity of information dissemination, leading to misinformation that can affect public perception and decision-making. Detecting fake news is critical to preserving the credibility of media sources, protecting democratic processes, and ensuring public safety. However, as Artificial Intelligence (AI)-driven Fake News Detection (FND) systems become more prevalent, ensuring AI safety is equally important to prevent unintended biases, ethical concerns, and adversarial exploitation of these models. Several challenges hinder the effectiveness of current FND models. Among these, cross-domain generalization and class imbalance are two critical problems that considerably impact the performance of detection systems. Although there are multiple challenges in FND, this thesis focuses on these two problems due to their widespread influence across various datasets and domains.
Cross-domain generalization is a complex problem because the characteristics of fake news differ significantly between domains. For example, fake news in politics may be presented in a completely different manner when compared to health-related misinformation. This variation in linguistic and contextual features makes it difficult for models trained in one domain to generalize to another. Still, important patterns can be present across domains and might be useful to detect fake news on different domains, helping with model generalizability. This thesis adopts a domain classification approach to address this issue, allowing the model to classify the domain before applying a domain-specific FND strategy, thereby improving overall generalization.
The class imbalance, where fake news is underrepresented in datasets compared to real news, presents another challenge, as models often become biased toward the majority class. This thesis addresses this problem by incorporating advanced Deep Learning (DL) techniques, such as Transfer Learning (TL) and class imbalance mitigation strategies, to ensure that models learn the nuances of detecting fake news even when instances are scarce.
The contributions of this thesis are categorized into the following areas: the first contribution is a comprehensive Systematic Literature Review (SLR) of the current DL techniques used in FND. This review highlights the challenges existing methods face, particularly with respect to cross-domain generalization and class imbalance, laying the foundation for developing novel approaches. The second contribution is developing a two-tiered, multi-domain FND framework. This framework integrates domain classification with FND, enhancing the model's ability to generalize across diverse domains, in addition to the mitigation of the imbalance issue by navigating various techniques. The third contribution is the introduction of a fine-tuned, prompt-tuned Llama 2 model. This model leverages pre-trained knowledge from Large Language Models (LLMs) and adapts it to the specific task of FND, improving performance by focusing on domain-specific characteristics. This thesis has an extensive evaluation of the proposed approaches using publicly available datasets from various domains, including politics, health, science, crime, and social media. The results demonstrate that our methods significantly improve the robustness and accuracy of the FND models, providing benchmarks for future research in this field. Additionally, the findings emphasize the importance of AI safety in FND, ensuring that models are not only effective but also fair, explainable, and resistant to adversarial manipulation. These contributions advance the state-of-the-art of FND and establish a more resilient framework for detecting fake news across multiple domains, addressing critical challenges that have previously limited the generalization and reliability of detection systems
Briser le silence : Disparitions et meurtres des femmes et des filles autochtones, représentation médiatique des violences : Analyse des images et des textes associés dans les médias autochtones
No full textLe silence entourant les disparitions et les meurtres des femmes et des filles autochtones révèle un système discriminatoire colonial entraînant un enchainement de violences à leur encontre. Marginalisées par les instances gouvernementales et judiciaires, les femmes et les filles autochtones sont invisibilisées dans les médias de grand public.
En mobilisant la théorie intersectionnelle, l'analyse fondée sur les médias autochtones, et plus particulièrement sur leurs presses numériques, apporte une nouvelle perspective de compréhension dans la représentation de la violence faite aux femmes et aux filles autochtones disparues et assassinées. En s'intéressant aux images médiatisées et aux textes associés, relevant des marches commémoratives pour les femmes du 14 février, les presses numériques autochtones se sont consacrées à dénoncer les violences issues du système discriminatoire colonial à l'appui d'une stratégie orientée vers la décolonisation