14589 research outputs found
Sort by
Chemistry and Computers, a Personal Recollection
Full text linkSubmitted versionEmployment as a senior chemical research engineer with a US oil company led me to a two year loan assignment to its refinery in Dart-mouth, Nova Scotia, in 1967-69. During this period, I came to like living in Nova Scotia, so in January, 1970, I resigned and became a computerization advisor to then President of Saint Mary’s University (SMU), Dr. Henry Labelle, S.J. Subsequently in 1972, I was offered the position at SMU as an associate professor of physical chemistry with a particular research interest in computerization of analytical chemistry procedures
Victim impact statements : investigating the factors involved in the delivery and perception of victim impact statements from self-reported victims of crime
Full text link1 online resource (165 pages) : graphsIncludes abstract and appendices.Includes bibliographical references (pages 109-122).Victim impact statements (VISs) are accounts from victims that describe the physical,
psychological, and financial harm from a crime. This research focuses on VISs in Canada,
specifically how observers perceive victims’ goals and credibility from their VIS. Study 1
interviewed real-life victims of crime via Zoom, which indicated that goals important to victims
included having their voices heard and being acknowledged. Study 2 asked 70 university
students to listen to three VIS audio-recordings from Study 1 to investigate goal congruence and credibility. The results identified a communication gap between observers and victims. Although adept at avoiding false positives, observers had low sensitivity (48%) for identifying the victims’ reported goals. Credibility ratings were found to be influenced by factors related to narrative structure and emotionality. The findings of this research have important implications for improving victims’ experiences and informing the Canadian criminal justice system of victims’ needs and objectives
Synthesis of cyclic polycationic allenes from a low coordinate carbocation
Full text link1 online resource (xxxi, 269 pages) : illustrations (some colour), charts (some colour), graphs (some colour)Includes abstract.Includes bibliographical references (pages 258-269).The preparation and reactivity of a 1,3-bis(dimethylamino)iodopropargyl salt (3‧2I) will be discussed. Prepared via a multistep pathway, the cations possess three potentially reactive functional groups, namely: (1) an alkyne, (2) a carbocation, and (3) an alkyl halide. This density of potential reactive sites provides a unique starting point for preparing heterocyclic compounds. The reactions of 3‧2I with organophosphorus compounds have been found to generate a variety of cyclic allenes, including the smallest in an eight-membered dicationic C3-P2C3‧2BPh4 to the largest in a twenty-two-membered tetracation [C6-P2C3]‧2I‧2BPh4.The incorporation of the allenic dication into a metallocene and into a chiral complex was also achieved. The newly isolated compounds have all been characterized using NMR and IR spectroscopy, LRMS, HRMS, and Xray crystallography
The inclusion-then-diversity concept : addressing hiring bias, inclusive business schools, and sustaining inclusion in organizations
Full text link1 online resource (194 pages) : charts, graphsIncludes abstract and appendices.Includes bibliographical references (pages 165-194).This dissertation examines the role of organizational inclusion as a foundational element in
fostering sustainable change in workplaces, education, and institutional systems. While diversity efforts are often prioritized, this research argues that without a strong foundation of inclusion, these initiatives may fail to create long-term impact. The Inclusion-Then-Diversity concept is introduced to emphasize that embedding inclusion first leads to more equitable policies, greater workforce retention, and enhanced organizational effectiveness. Using a manuscript-based format, this research explores inclusion across three key areas. Chapter 2 investigates hiring discrimination against Indigenous men with criminal records, highlighting the limitations of conventional fairness-based policies and advocating for culturally responsive hiring practices. Chapter 3 shifts focus to business education, analyzing how inclusive learning environments, particularly through co-curricular initiatives, can enhance student belonging and success. Chapter 4 introduces a framework for sustaining inclusion within organizations, offering a model for embedding inclusive policies into long-term structural change. Collectively, these chapters illustrate that inclusion must be deliberately integrated into policies, institutional cultures, and decision-making processes to create workplaces and educational environments that genuinely support and value all individuals.
This research contributes to the growing discourse on organizational inclusion by providing
empirical and conceptual insights into how inclusion can be effectively operationalized across
different sectors. By prioritizing inclusion before diversity, organizations can cultivate
environments where diverse employees and students not only enter but thrive. The findings have implications for policymakers, business leaders, and educators seeking to build more inclusive and equitable institutions
The era of high-resolution and broadband x-ray spectroscopy : insights into collisional and photoionized plasma in AGN
No full text1 online resource (xiii, 224 pages) : illustrations (some colour), charts (some colour), graphs (some colour)Includes abstract and appendix.Includes bibliographical references (pages 201-224).High-resolution X-ray spectroscopy provides a powerful tool for probing the physical
conditions of the environments surrounding active galactic nuclei (AGN). The soft X-ray
band, in particular, is often populated by spectral features originating from either emission
or absorption of ionized plasmas. The ionization mechanism of these plasmas, either colli-
sionally or photoionized, along with the physical conditions of the emitting region, will leave
an imprint on the intrinsic spectrum. Discerning between the two ionization mechanisms,
along with the physical conditions of the emitting plasma, is paramount for understanding
how AGN influence their host galaxy. To explore these ionized plasmas, we examine the
high-resolution spectra of two Seyfert galaxies, Mrk 1239 and NGC 1068, to gain insights into
where, why and how collisionally and photoionized plasmas appear in AGN spectra. This
work utilizes new and archival data from XMM-Newton, Chandra, NuSTAR, and Swift, and
incorporates both broadband spectral and timing analysis alongside high-resolution spec-
troscopy. The collisional plasma in Mrk 1239 was initially attributed to star formation in
the host galaxy, but further analysis suggests it is more consistent with material crashing
into the interstellar medium (ISM), driven by either a young jet or radiation pressure. We
propose that the unique properties of Mrk 1239 could be an indication of feedback early in
the AGN duty cycle. The photoionized plasma in NGC 1068 was examined to explore the
utility of the Fe xxii emission doublet as a density diagnostic in AGN. Utilizing the spectral
fitting software spex, we demonstrate that density estimation is highly dependent on input
parameters, such as ionizing continuum and ionization state. This analysis enabled us to
gain insights into the systematic uncertainties inherent in our methodology
Vocational interests in the CAF
No full text1 online resource (120 pages) : colour graphsIncludes abstract and appendices.Includes bibliographical references (pages 80-93).The Canadian Armed Forces’s (CAF) recruitment and retention crisis begs a greater
understanding of the people within the organization and their experiences. Alignment between employees and their jobs influence attitudes, behaviours, and well-being. Vocational interests are an area of individual difference that can be leveraged to improve this alignment. Serving CAF members (n = 883) completed a survey including the O*NET Interest Profiler and self-report measures of fit, commitment, satisfaction, anxiety, engagement, and OCB. Latent Profile Analysis revealed four interest profiles. The Social Strategist profile was associated with the most positive work outcomes. Regression analyses provided little support for the notion that congruence is associated with positive work outcomes. Congruence between an individual’s interests and their occupation was the strongest predictor of work outcomes. Understanding the interests of members and their relationships with work outcomes can inform CAF policies and strategies to improve members’ experiences and inform recruitment efforts
Artificial minds : can AI ace psychopathy tests used in selection?
Full text link1 online resource (74 pages)Includes abstract and appendices.Includes bibliographical references (pages 50-58).The increased accessibility of Large Language Models (LLMs) poses an unprecedented
threat to selection processes as more individuals are using Artificial Intelligence (AI). Job
candidates often misrepresent themselves during the job selection process, including on
personality measures. One personality trait of interest is psychopathy, due to the detrimental
organizational outcomes that are associated with high psychopathy in the workplace. Thus, it is important to understand how AI impacts applicants’ ability to fake different measures of
psychopathy, including a recently validated implicit test (Conditional Reasoning Test of
Workplace Psychopathy). Select LLMs and tests of psychopathy were analyzed using multiple ANOVAs across prompting conditions to determine the success/failure of LLMs in faking various tests of psychopathy. Results reveal select LLMs can distort their scores on psychopathy measures, across conditions. Differences between LLMs were also found, with some outperforming others. These results add to the growing body of literature on AI and faking
2025-03-14 Senate Minutes and Agendas
No full textMinutes and Agenda for the March 14, 2025 meeting of Saint Mary’s University Senate
Transformer based approach to monocular depth estimation : adapting Depth Anything V2 for inventory management in retail
Full text link1 online resource (vi, 80 pages) : illustrations (some colour), charts (some colour), graphsIncludes abstract.Includes bibliographical references (pages 76-80).This thesis presents a comprehensive study on adapting Depth Anything V2, a state-of-the-art monocular depth estimation model, for retail environment applications. This work addresses the unique challenges of depth estimation in retail settings, including densely packed shelves, reflective surfaces, varying lighting conditions, and frequent occlusions. This research developed a data collection methodology using Intel RealSense D435 cameras in a controlled laboratory environment simulating retail conditions. The dataset comprises high-quality RGB-D pairs captured across systematic variations in camera positions, lighting conditions, and product arrangements. The implementation leverages Depth Anything V2's Vision Transformer backbone, fine-tuned on the custom retail dataset. The experiments demonstrate significant improvements through domain-specific adaptation. The study evaluated three Vision Transformer variants achieving δ₁ accuracy exceeding 97% across all models. The ViT-B variant emerged as optimal, achieving the lowest absolute relative error and competitive RMSE. Compared to baseline models, retail-adapted versions achieving up to 63.95% RMSE reduction
Artificial intelligent powered fraud detection in banking : innovations, challenges and preventive strategies
No full text1 online resource (iii, 43 pages)Includes abstract.Includes bibliographical references (pages 39-43).As artificial intelligence (AI) becomes integrated in the banking industry, it enhances
efficiency and introduces vulnerabilities while AI-driven frauds have emerged as a significant threat. This study explores the risks of AI-enabled frauds and strategies for prevention. Through literature review, secondary data analysis, and examples, this thesis identifies key fraud types, including generative AI’s role in creating deepfakes for signatures, videos, and voice impersonations, which cause financial losses and undermine trust.The thesis identifies cutting-edge countermeasures to these risks, including shared large language models (LLMs), automation tools, liveness testing, compliance controls, and machine learning-based anti-fraud technology. These methods strengthen data security and enhance fraud detection. The results highlight the critical necessity for flexible tactics to counteract changing fraud schemes and provide helpful advice for bolstering financial defenses and guaranteeing the stability of banking institutions in a time of swift technological advancement