82837 research outputs found
Sort by
An Investigation into the Expression and Role of TSLP During the Anti-Viral Response of Airway Epithelial Cells
No full textThe airway epithelium is one of the first lines of defence against airborne pathogens. Airway epithelial cells are crucial in initiating and directing the subsequent immune response to deal with invading pathogens rapidly. An early mediator produced by epithelial cells is the alarmin cytokine, thymic stromal lymphopoietin (TSLP), which is known to activate several immune cells. However, overexpression of TSLP is the underlying cause of several diseases, including asthma and allergy, and contributes to increased pathophysiology of other respiratory viruses. Little is known regarding the role of the airway epithelium and TSLP during SARS- CoV-2 infections, the virus responsible for COVID-19. Additionally, the role of TSLP autocrine/paracrine signalling within the epithelium is poorly understood.
We investigated the expression of TSLP in COVID-19 patients at the University of Alberta hospital and found that they had a trend towards elevated systemic TSLP, although this was not statistically significant, which correlated with increased hospitalization duration. Several other cytokines were measured, and significant plasma IL-15 and CXCL10/IP-10 increases were detected. In cultured bronchial epithelial cells from healthy donors, we measured significant increases in intracellular and supernatant TSLP expression in response to SARS-CoV-2 infection but not in nasal or gut epithelial cells.
To increase our understanding of the role TSLP has within the epithelium, we created a simple model to virally induce epithelial-derived TSLP production and block it using TSLP- neutralizing antibodies. We confirmed that the viral mimetic poly I:C can induce a strong TSLP response through a TLR3-specific mechanism. Furthermore, we characterized the expression of several key antiviral and inflammatory genes, including TSLP, TSLPR, RIG-I, MDA5, IL-25, IL-33, TL1A) and interferons. For the first time, we observed that TL1A is significantly
I
upregulated in airway epithelial cells in response to poly I:C treatment. Additionally, we made a novel observation showing that epithelial-derived TSLP may contribute to regulating antiviral receptors such as RIG-I and MDA5.
Our goal is that the data presented in this thesis sheds light on the vital role of the airway epithelium and TSLP during COVID-19 infections and the broader context of antiviral defences. Our findings are intended to increase our understanding and appreciation for the important role of epithelial cells during viral infections
Online Conversion under Horizon Uncertainty: From Competitive Analysis to Learning-Augmented Algorithms
No full textOnline allocation problems involve making sequential decisions under incomplete information, where inputs are revealed incrementally over time. A prominent subclass of online allocation problems is the one-way online conversion problem (a.k.a. one-way trading), which focuses on selling (or buying) a single type of divisible resource under dynamically changing prices. Decision-making in online conversion requires balancing immediate reward against the potential for better future opportunities. A critical yet less-explored aspect of online conversion is the impact of time horizon uncertainty, which determines the duration over which decisions are made. The horizon may be predetermined, revealed partway, or entirely unknown, introducing layers of complexity that significantly influence conversion strategies. Additionally, practical constraints such as box constraints, which limit the maximum allowable trade per step, further complicate the decision-making process and demand more nuanced algorithmic approaches.
Despite progress in addressing online conversion problems, significant research gaps remain. Existing studies often focus on unconstrained settings or assume complete knowledge of the horizon. Few works explore the combined effects of horizon uncertainty and practical constraints like box constraints on algorithm design and performance. Additionally, leveraging horizon predictions to enhance performance has been underexplored in this context. This thesis addresses these gaps by making two main contributions. First, we propose a unified algorithm to address three models of horizon uncertainty—known horizon, notification at a specific step, and unknown horizon—under both constrained and unconstrained settings. Through competitive analysis, the algorithm achieves tight guarantees, demonstrating optimal performance across all scenarios. Second, the thesis extends the unified algorithm to incorporate horizon predictions, introducing a learning-augmented algorithm that bridges the gap between worst-case and average-case performance. By balancing robustness under adversarial conditions with consistency when predictions are accurate, the algorithm exhibits adaptability in uncertain environments. Together, these contributions advance the theoretical foundation of online conversion and provide practical insights for applications where horizon uncertainty and constraints like box constraints play a critical role
A Novel Cost Minimizing Strategy for Cooperative Relay and Edge Computing-based Wind Power Communication Network
No full textWind power frequency regulation relies on reliable communication between wind farms and power dispatch center (PDC), which is critical for ensuring the accuracy of frequency regulation. However, data transmission errors and delays may lead to deviations in control commands, increasing the reliance on ancillary services and thereby raising the operational costs of PDC. Therefore, this paper proposes a communication-enhancing strategy that integrates cooperative relay and edge computing (CRE) to improve communication reliability. Edge node (EN) devices with computational capabilities are employed as relays to enable data preprocessing at edge-side. Furthermore, a bandwidth compression model is developed based on Amdahl’s law, hardware constraints of EN, and thermal wall effect, to characterize the coupling between computing power and bandwidth and support efficient resource allocation. On this basis, a cost model incorporating packet loss, frequency regulation response time, and control errors is established, and the cost minimization problem is formulated as a Stackelberg game to jointly optimize resource allocation and pricing strategy. Simulation results demonstrate that the proposed strategy improves resource
utilization efficiency. Compared with benchmark schemes, it reduces the total cost of PDC by 33.82% and increases the profit of EN by 21.53%, while exhibiting strong scalability across different wind farm scales
An Administrative Leader’s Perspective on Balancing Change Leadership and Change Management in Higher Education
No full textThis capstone paper explores the case of a schedule modernization initiative at the University of Alberta. The initiative is preparing to enter design and implementation with the intention of deploying new technology, along with possible changes to policy, process, and work allocation. Given the importance of the class schedule to the operations of the University, a change of this nature will require significant change leadership and change management. The case will be explored through Bolman and Deal’s Reframing Organizations: Artistry, Choice, and Leadership (2021) and Kotter’s eight step change management model (Bolman & Deal, 2021, pp. 405-407; Kotter, 2012, 2017; Galli, 2018, pp. 126-127) with the intention on assessing the potential of interfacing the two frameworks to surface to organizational considerations that may require focused change leadership and change management. In doing so, I hope to hone my efforts as an administrative leader within the organization as I prepare to help lead the change
Characterization of Drilling Waste Sumps in Northern Alberta, Canada
No full textSumps, or earthen pits, are often used to store and dispose of solid and liquid drilling
wastes from oil and gas drilling operations; the contents of these sumps can vary depending
on the region and the depth of the related drilling project(s). The variation includes chemical
composition of the wastes found in sumps, including organic compounds (i.e.:
hydrocarbons), salts and metals, and presents potential environmental and health risks. It is
difficult to estimate the number of un-reclaimed sumps present in Alberta as gaps in recordkeeping practices increase the further back in time one looks; the magnitude of the ‘sumpproblem’ is not well understood. Although generally undocumented early on, drilling wastes
were often characterized before deposition into an earthen-sump and records were kept by
the energy producer. To date, little public or academic literature has been published on the
characterization or contents of sumps, the transport of constituents from sumps, or the
environmental and health risks they may pose. This thesis began with broad characterization
of the sump material in both north central Alberta and northeastern British Columbia. In this
thesis, 5 sumps were sampled and analysed to characterize the geochemistry of the geologic
media at the three different depths of the sump: above, within and below the drilling waste
mix zone. A variety of experiments were conducted, including standard commercial lab
testing for the initial concentrations of BTEX and PAHs, batch leaching and sedimentpacked flowing column experiments (for BTEX), total organic carbon (TOC), x-ray
diffraction (XRD), sequential extraction for metals in the leaching experiments and a brief
microbial characterization experiment. Initial concentrations for the sump sites studied in
this thesis varied: benzene, ethylbenzene, total xylenes and in instances lighter-end (F1)
hydrocarbons were found to be above Alberta Tier 1 guidelines, while toluene was not. The
hydrocarbons of interest are those that are included in standard environmental testing and
considered highly mobile because they are lighter-end hydrocarbons, such as the volatile
organic compounds (VOCs) benzene, toluene, ethylbenzene, xylenes (BTEX) and F1
hydrocarbons. Rudimentary calculations were made on the likelihood of BTEX contaminant
transport, the calculations suggest that BTEX is unlikely to be transported from the sump
into the surrounding environment. The results of this thesis suggest that sumps of similar
characterization to those studied in this thesis imply that even though the initial
concentrations of BTEX may be over acceptable standard guidelines the likelihood of
contamination leaching from the sump is possible. The findings of this thesis suggests that
many sumps may leach sump material and therefore could pose a risk to the surrounding
environment
GrEx-Budget Section
No full textTips for what to include in your SSHRC budget, strategies for writing a persuasive budget, sample IG and IDG budget templates, and select university guidelines
OROFACIAL CLEFTS IN ALBERTA: DOES SOCIOECONOMIC STATUS PLAY A ROLE?
No full textOrofacial clefts (OFC) are a common congenital anomaly, affecting 1 per 1,000 to 1 per 2,000 total births depending on the type (Cleft Lip, with or without Cleft Palate and isolated Cleft Palate). Among other etiological factors, the low socioeconomic status (SES) index has been associated with an increased risk of OFC in the US. However, an Ontario population-based study found no association between neighbourhood income and OFC risk. In Alberta, to the best of our knowledge, data on the association of maternal SES and OFC is limited.
To evaluate the incidence and geographic distribution of OFCs in Alberta and their association with maternal SES, we did a population-based cross-sectional study of all live births born to Alberta-resident mothers from January 1, 2006- December 31, 2019, with or without OFC, using data from the Alberta Congenital Anomalies Surveillance System, Data and Analytics Department of Alberta Health Services and Alberta Perinatal Health Program. Maternal demographics, pre-pregnancy health conditions, lifestyle data, perinatal and neonatal variables were collected for OFC cases, identified through ICD-10 codes. Maternal residence Forward Sortation Area (FSA) data and Pampalon Material Deprivation Index (MDI) and Social Deprivation Index (SDI) assigned to all live births and OFC cases were abstracted. The MDI and SDI were used as proxy indicators for maternal SES. MDI and SDI quintiles were categorized into Least deprived (Quintiles 1 and 2), Moderately deprived (Quintile 3) and Most deprived (Quintiles 4 and 5). Statistical analysis focused on Univariate analysis rather than Multivariate analysis because potential confounding variables, such as maternal smoking and alcohol use, were unavailable for all live births.
After excluding duplicate birth records and non-Alberta resident mothers from a total of 745,341 live births in Alberta, there were 717,635 live births born to Alberta-resident mothers. Pampalon indices were available for 215 OFC cases and 683150 (95%) live births (Urban = 566,808; Rural = 116,342). Provincially, the highest number of live births occurred in the most deprived quintiles of MDI (43%) and SDI (45%). There were 224 OFC cases provincially (0.31 per 1,000 live births, 47% male); 180 OFC cases (80%) in urban Alberta, especially in large metropolitan cities (Edmonton, n = 48 and Calgary, n = 58), and 44 (20%) OFC cases in rural Alberta. The OFC rate was higher in rural Alberta (0.38 per 1,000 live births) than in urban Alberta (0.30 per 1,000 live births). The highest proportion of OFC cases in all of Alberta were noted in high levels of deprivation (MDI 46% and SDI 47%). The most deprived FSAs in large metropolitan cities had OFC rates around 1.5 times higher than less deprived areas. In rural Alberta, the highest OFC rate (1.17 per 1,000 live births) was observed among moderately materially deprived and less socially deprived mothers. Univariate analysis of OFC data from all of Alberta showed younger maternal age, higher rates of maternal drug and alcohol use and lower birth weight among OFC cases in the most deprived MDI quintiles. A higher non-significant rate of maternal smoking was noted in most deprived MDI (20%) and SDI (19%). In all of Alberta, a non-significant unadjusted Odds Ratio (OR) for OFC cases at all levels of MDI and SDI was noted. Similarly, non-significant OR was found in large metropolitan cities for the most deprived quintiles of MDI and SDI. Both metropolitan cities observed a non-significant moderate positive correlation with material deprivation (rs = 0.50, p=0.637) and a significant strong positive correlation with social deprivation (rs = 1.00, p=0.000).
A higher proportion of OFC cases were born to urban Alberta mothers who were both materially and socially deprived; rural Alberta mothers of OFC cases had most material but not social deprivation. Unlike other population studies, there was no significant association between maternal SES and the development of OFC, which could be due to methodological issues that may not have shown the true measure of association between OFC and SES in Alberta. Prospective population studies including OFC cases among total births, with a larger sample size exploring the urban/rural distribution of OFC cases among various levels of material and social deprivation could yield valid and reliable results in Alberta
Continual Preference-based Reinforcement Learning with Hypernetworks
No full textReinforcement learning (RL) has seen many applications across control problems. Central to all these problems is the reward. However, specifying reward functions accurately aligned with the user preferences is often difficult. A potential solution is preference-based RL that learns a policy and a reward function from preferences over behaviors given by a teacher (e.g., a human). Existing works assume the teacher’s model of preferences (underlying reward function) to be static. However, a human-agent interaction in the real world is likely non-stationary due to the human’s changing desiderata and beliefs over time. Hence, we propose a continual preference-based RL setting involving continual learning of reward functions based on the non-stationary teacher preferences. The setting’s (often conflicting) dual objectives are to learn new knowledge and preserve old knowledge. To address these, we propose using regularized hypernetworks that achieve both objectives and learn reward functions well aligned with the teacher’s preferences. Hypernetworks are neural networks outputting the parameters of a target model (i.e., reward function), given a task-specific input-context. The tasks differ in terms of the reward function, implied by the preference model. We perform training on each new task using a regularizer on the input-output mappings of previous tasks to learn task-specific models summarized in a single shared hypernetwork. We demonstrate the efficacy of our method against several continual learning baselines on two continuous control domains. Lastly, we analyze the behavior of continually learned reward functions by varying the amount of feedback from the teacher. We show how reward overfitting can occur in our setting, the associated risks, and how to prevent it. Altogether, this work takes a step toward more realistic applications involving a continual and non-stationary interaction between humans and agents
Understanding Mothers’ Experiences of Maternal Guilt and Maternal Shame and How These Experiences are Expressed in the Mother-Child Relationship
No full textThe distinct and intense emotional experiences of shame and guilt may accompany mothers’ experiences in their role as mothers. Although experiences of maternal guilt have been studied, less is known about experiences of maternal shame and even less is known about how these emotional experiences may be expressed in the mother-child relationship. This current doctoral research aimed to (a) better understand mothers’ experiences of maternal guilt and/or shame and (b) explore how these experiences may be expressed in the mother-child relationship. A qualitative descriptive study was conducted with four mothers via semi-structured virtual interviews. Through thematic data analysis, nine themes and two subthemes were identified: (a) ups and downs of being a mother, (b) expectations of mothers (with two subthemes: 1. internal expectations and 2. external expectations), (c) mothering ideologies and practices, (d) responsibilities of caregiving and child development, (e) experiences of maternal guilt, (f) experiences of maternal shame, (g) expressions of maternal guilt and shame, (h) emotions experienced in motherhood, and (i) talking about experiences of mothering. Clinical significance, limitations, and future research will also be discussed
Towards Smart Textiles Produced from Electrospun PVDF-HFP and its Composites
Get PDFAs exploring new energy harvesting methods is gaining momentum as a solution to the current environmental crisis, conversion from mechanical energy to electrical power in a green, sustainable fashion has been comprehensively assessed. Piezoelectric materials offer a scalable mechanism for energy conversion via intrinsic polarization properties. Advancements in the manufacturing of nanoscale materials have enabled the production of piezoelectric materials with high flexibility, enhanced piezoelectric performance, and increased versatility in applications. Polyvinylidene fluoride-co-hexafluoropropylene (PVDF-HFP) is one such piezoelectric material with immense potential for sensing and energy harvesting applications, such as smart textiles, pressure sensors, or biomedical sensors. The incorporation of cellulose nanocrystals (CNC) in PVDF-HFP presents a promising route to improve its piezoelectric properties. This thesis investigates the enhancement of the piezoelectric properties of PVDF-HFP through CNC reinforcement, explores applications in sensing and drug delivery, and models the manufacturing process for better tailorability of the piezoelectric properties.
The thesis starts with a literature review accompanied by machine learning to predict the piezoelectric output of electrospun polyvinylidene fluoride (PVDF) mats with various fillers under mechanical stress or strain. PVDF was selected for its similar piezoelectric properties to PVDF-HFP and the large quantity of related research. This work provided a predictive framework that could be extended to PVDF-HFP, demonstrating that enhancing the β-phase content was the primary strategy for improving piezoelectric performance. Specifically, it highlighted that incorporating nanofillers could increase the voltage output of electrospun PVDF-based polymer fibers with a classification model employed to categorize output voltage ranges based on practical electronic applications.
From the literature review and machine learning model, nanofillers were found effective in improving the piezoelectric properties of PVDF-based polymers. Thus, the next study focused on electrospinning of PVDF-HFP nanofiber mats reinforced with CNC. The incorporation of CNC significantly enhanced the β-phase content of PVDF-HFP and its piezoelectric output. The study systematically varied CNC concentrations and other experimental parameters to determine the optimal conditions for β-phase formation and mechanical reinforcement. The application of the PVDF-HFP/CNC mat as a pressure sensing device that could be employed as a sensor for detecting carpal tunnel syndrome for pianists was achieved.
Based on the optimized conditions established for random electrospinning, the investigation extended to yarn electrospinning of PVDF-HFP/CNC. Yarn electrospinning yielded twisted continuous yarns. PVDF-HFP yarns offered great uniformity and mechanical properties, as well as significantly enhanced piezoelectric performance compared to randomly oriented nanofiber mats. One application explored in this thesis was the use of these yarns in touchscreen gloves, providing real-time motion-sensing capabilities. This innovation can advance human-device interaction while offering health monitoring benefits.
The application of PVDF-HFP/CNC was explored beyond its piezoelectric properties in the thesis by investigating its potential for drug delivery. PVDF-HFP/CNC yarns exhibited pH-responsive drug release performance with sustained structural stability at pH and temperature environments suitable for wound healing. Their drug release behaviors conformed to the Ritger-Peppas model, suggesting Fickian diffusion dominated the release kinetics. After loading with levofloxacin, PVDF-HFP/CNC yarns exhibited strong antimicrobial activity and cytocompatibility, emphasizing their potential for advancing antibacterial wound healing in suture applications.
To further understand how the manufacturing method affected the properties of yarn electrospun products, computational fluid dynamics (CFD) models were constructed for the random and yarn electrospinning processes. Taylor cone formation and jet trajectories in stable and unstable regions were simulated to understand yarn formation mechanisms. This study provided critical insights into how the simultaneous presence of two electric fields promoted jet collision during electrospinning, which was never investigated in the literature before. The CFD model, subjected to future refinements, can serve as a predictive tool for manufacturing.
This study was the first to utilize machine learning for predicting the piezoelectric performance of electrospun PVDF-baed polymers and their composites. Yarn electrospinning of PVDF-HFP/CNC was introduced for the first time, with novel applications explored in motion-sensing touchscreen gloves and drug-loaded sutures. The theoretical simulation and analysis revealed the underlying mechanisms of random and yarn electrospinning, providing a deeper understanding of the process. The findings offer valuable insights into developing next-generation smart textiles and medical devices, highlighting the versatile and impactful role of PVDF-HFP/CNC composites in emerging technologies