21793 research outputs found
Sort by
Towards a Sociology of Machines: Reimagining Human-Machine Relations
Get PDFThe dissertation proposes the outlines of a sociology of machines for understanding human-machine relations, especially those that exceed the dominant normative frameworks. Bringing together social theory, artificial intelligence research, and human-machine interaction literature, the dissertation argues for the need to devote attention to alternative realities with technologies from a sociological perspective. This move serves as a critique of the dominant manner in which technologies are thought with, which is as means to the demands of an instrumental rationality. In the face of threats to other forms of life and relationality, the thesis insists on doing things otherwise, including sociological conduct itself. The thesis proposes using creative methodologies such as research-creation as a way to reinvigorate sociology to develop its capacity to address the multiplicity of human machine relations.
The thesis centers on machines that cannot be easily sublimated under frameworks of instrumentality, control, and management. It looks at relations with useless machines, broken machines, or machines that are treated as legitimate social actors. This allows the outlines of a theory to emerge based on the recognition of multiplicity of machines as well as humans. Thus, while establishing itself on an alternative ontology that takes seriously the contributions of other-than-humans in constitution of social reality, it also moves beyond the view that machines simply extend the logic of the contemporary power structures.
The thesis addresses these questions first by interrogating the conception of agency at the intersection of George Herbert Mead’s and Alan Turing’s thought. Secondly, it investigates the intimate relations developed with social robots to understand the development of different sensibilities in humans. Then it inquires into generative models to better gauge the necessary transformation of our conceptions of agency. Finally, a diagnosis is made through Heidegger’s thought on technology to grasp the essence of technology, and a pathway is deduced for a sociology of machines that employs lively methodologies to make sense of the contemporary moment of machinic socialities
Prioritizing and Mitigating the Supply Chain Risks of a Dairy Company in an Unstable Economic Environment
Get PDFThe aim of this paper is to implement the Failure Mode and Effect Analysis (FMEA) approach to assess, mitigate, and prioritize risks in the dairy industry, with a particular emphasis on a specific company. The values of occurrence (O), severity (S) and detection (D) of failures will be determined. These estimated values will be used to calculate a risk priority number (RPN), for each potential failure, ultimately allowing to us to develop a well-defined and comprehensive risk prioritization list. Through this paper, we apply a practical and refined FMEA approach, tailored specifically for the dairy industry. While our findings draws on commonalities shared with other dairy companies all around the world, it is necessary to recognize that each dairy business has distinct characteristics and challenges. Since this case study focuses on a particular dairy company in Lebanon, a country that is currently in the midst of an economic crisis, we gain valuable insights into the unique risks and corrective actions undertaken by the company. Moreover, this paper presents an innovative and enhanced version of the traditional FMEA approach by incorporating two additional factors, the application rate and the weight-based model. While the application rate (A) enriches the FMEA analysis by assessing how effectively and efficiently the studied company applies the corrective measures, the weight-based model enables a more comprehensive evaluation of the analyzed risks by taking into consideration the relative collective expertise of members involved in the supply chain. Once the most substantial threats were identified, we were able to provide and improve risk mitigation strategies for those threats. Eventually, by improving the firm’s decision-making on risks, we intend to minimize potential future losses associated with specific scenarios. Moreover, future research
paths are highlighted as a result of this study
Visualization, Quantification, And Analysis Of Inter-rater Variability To Enhance Deep Learning-based Medical Image Segmentation Of Paraspinal Muscles
Get PDFDeep learning-based medical image segmentation has revolutionized healthcare diagnostics. While the accuracy offered by these models is important, ensuring their practical implementation requires a comprehensive reliability assessment. Among the sources of uncertainty, inter-rater variability, which reflects natural disagreements among annotators, has been historically overlooked. Studying this variability can be a key factor in improving the robustness and reliability of models.
Basing our experiments on paraspinal muscle segmentation, which has significant value in studies related to low back pain, in this dissertation, we first proposed a novel multi-task TransUNet model to accurately segment paraspinal muscles while predicting inter- rater labeling variability visualized using a variance map of raters’ annotations. Benefiting from the transformer mechanism and convolution neural networks, our algorithm was shown to perform better or similar to the state-of-the-art methods while predicting and visualizing multi-rater annotation variance per muscle group in an intuitive manner. Subsequently, we studied the relationship between inter-rater variability and aleatoric/epistemic uncertainties, in the context of DL model architecture and label fusion methods. Specifically, we measured aleatoric and epistemic uncertainties using test-time augmentation, test-time dropout, and deep ensemble to explore their relationship with inter-rater variability. Furthermore, we compared UNet and TransUNet to study the impacts of Transformers on model uncertainty with two label fusion strategies. This thesis provides novel frameworks for visualizing, understanding and quantifying inter-rater variability to better inform relevant deployment and implementation of DL models for medical image segmentation
Echoing the Cicadas: Neo-Sophism, Gamification of Knowledge & the Expertise-Illusion in the Contemporary Infosphere
Get PDFFor Plato, writing’s power for sophistry and its authority over the reader is founded on writing’s nature as incomplete without the embodied presence of the author, creating an asymmetrical reader-text relationship wherein the reader is active, and the text passive. The modern infosphere presents a new problem but at the same time connects back to this Platonic issue. The relationship between agency and patiency in Plato’s world of textuality is transformed by the mechanisms of the contemporary infosphere. Instead of silence, modern readers are confronted with a bombardment of voices competing for their attention. Conceived as a new tradition, the infosphere presents readers with an unbounded and ever-shifting horizon of texts and references as already interpreted answers. From this unbounded horizon—a product of what I call the collective memory paradox—the expertise-illusion arises within the consciousness of certain types of readers, when the dialectic between question and answer is dissolved through the isolating character of textuality, which is intensified by its form in the internet age. Genuine communication is corrupted when the reader’s agency is unknowingly replaced by the info-sophist’s agency. This is exemplified and amplified through various forms of the ‘gamification’ of knowledge, which encourages readers to seek compulsive accumulation of information, diverting them from thinking their own way to knowledge. This gamification can provide a false sense of agency, making readers think they ‘have’ knowledge, and thereby expertise. However, they have merely won the game of acquiring it within a framework that provisions them to do so
Animer des grands groupes où chaque participant⋅e contribue à l'émergence d'un projet commun
Get PDFOn retrouve souvent deux dynamiques dans les grands groupes. D'une part les conférences, forums, congrès, symposiums, conventions ; d'autre part les hackathons, forum ouverts, open cafés, unconferences, game jams. Malheureusement, ces dynamiques ne favorisent pas l’émergence d’un projet commun. Une troisième voie permet de co-créer, à partir des réalités individuelles, un contenu commun. L’engagement de chacun⋅e permet d’atteindre un résultat partagé, en préservant l’énergie unie. Ce document présente six étapes pour mettre en place une telle dynamique
Evaluating Occupational Exoskeleton Performance in Construction: Towards Standard Testing for the Construction Industry
Get PDFOccupational exoskeletons (OEs) are wearable devices designed to as-sist and enable human motion for workers in industries ranging from manufacturing to construction. As with workers in other industries, the opportunity pro-vided by OEs for construction workers is to reduce injury rates for the benefit of worker health and productivity. Potential risks also exist, including discomfort, compromised balance, snags, and increased stress in the unassisted regions of the body. The challenge comes in finding effective OEs for specific construction trades working on specific project types. To meet this challenge, this paper aims to develop a standardized OE efficacy evaluation framework for passive back-support exoskeletons (BSEs) for rebar workers. In-lab efficacy evaluation can lead to in-field effectiveness evaluation giving evidence for practical OE regulations, guidelines, and ergonomic risk indices to be used by the construction industry. The evaluation framework will include the assessment of a BSE’s effects on safety, productivity, and acceptability, including a tool for estimating an OE’s return on investment. The framework will allow for sensitivity analyses for hu-man attributes, including gender, age, and workers with prior injuries. It is hoped that this work will provide an integral step towards large-scale OE adoption for the construction trades
Student Motivation Matters: A Longitudinal, Person-Centered Perspective on Student Achievement Goals and Psychosocial Functioning
Get PDFThe achievement goals motivating students’ academic engagement share important associations with their academic and socioemotional functioning. However, the literature on achievement goals remains mainly cross-sectional and variable-centered, ignoring how these goals act in combination over time. This thesis provides a longitudinal person-centered examination of the relations between students’ achievement goals and their academic and socioemotional functioning at different educational stages.
In Chapter 2, we relied on Latent Profile and Latent Transition Analyses among a sample of 619 French Canadian elementary school students to estimate profiles encompassing students’ own achievement goals and their perceptions of the goals held for them by their parents and teachers. Four profiles emerged which remained stable from one school year to the next: Low on all Goals, High on all Goals, Mastery-Oriented, and Low Mastery Goals. Students were moderately to highly likely to remain in their profile over time, and all profiles reflected alignment between students’ personal goals and their perceived caregiver goals. Higher perceived competence was associated with higher mastery profiles, and students corresponding to the Mastery-Oriented profile fared best in terms of anxiety and achievement.
In Chapter 3, we relied on Growth Mixture Analyses to examine the developmental heterogeneity of burnout amongst 513 Finnish students transitioning from upper secondary to post secondary school. Four trajectory profiles emerged: High and Decreasing, Moderate and Decreasing, Low and Increasing, and Low and Stable. High initial levels of global self-esteem and mastery-extrinsic goals were associated with the Low and Stable profile, whereas high performance goals were associated with the High and Decreasing profile. The Low and Stable profile protected students from school dropout, underachievement, and substance use, whereas the High and Decreasing profile was associated with lower achievement, and more substance use.
Taken together, our results offer novel and nuanced person-centered insight into the relations between students’ achievement goals and various markers of academic and socioemotional functioning over the course of the educational trajectory. The results from this thesis can be used to inform developmentally relevant interventions seeking to enhance students’ mastery goals, perceived competence, and self-concept to promote their psychosocial functioning over time
Starting in the Homes: Patriarchy, Power, & Well-Being Addressing Intimate Gendered Violence Against South Asian Women in Montreal
Get PDFSouth Asian women’s organizations (SAWOs) in the United States and Canada have reported high prevalences of intimate gendered violence in their communities for several decades now despite a lacune in both research and attention. This project takes as its starting point the following question: what factors are involved in addressing intimate violence against South Asian women in Montreal and in supporting their well-being? Research methods are based on critical ethnography, writing against culture, and feminist anthropological frameworks and include participant observation, textual analysis, and interviews with front-line workers at and with members of the South Asian Women’s Community Centre of Montreal (SAWCC). This thesis is the start of a larger work which uses the analytical tool of mapping. This mapping begins in women’s homes, household and home country, looking at the patriarchal gender norms and ideologies which women are socialized into that maintain patriarchal power which are reproduced through key institutions of the family and marriage. Next, four major taboos, i.e. divorce, sex, mental health, and violence, are discussed to provide insight into how they shape women’s experiences and responses to violence and serve to keep the status quo of patriarchal power. Finally, interventions into these factors by SAWCC are discussed as possible methods for “pulling the roots” of women’s subordination through empowering South Asian women and changing norms. Lastly, the next steps for this analysis are named which entail the mapping out of remaining topographies of key factors that intersect and compound with those named in this thesis. This research found that addressing intimate gendered violence and supporting the well-being of South Asian women requires a critical understanding of specific realities at the intersection of ideological and structural factors, the role of (mis)understandings and use of the concept of culture in ‘fields of power’, integration of anti-racism into anti-violence work and services, and the recognition of the crucial work done by community and grassroots organizations that provide key interventions
The Sonic Operative: Sound, Activism, and the Politics of Response to Post-Soviet Crises
Get PDFThe sonic operative is a propositional formulation of what the dissertation outlines as a situated, implicated, and contextually grounded work of sound by which to craft activist response. I come up with this formulation to think with possibilities emerging in practices invested in a sonic format of justice seeking action. Using the post-Soviet contexts of dictatorship, imperialism, colonialism, capitalism, and ecocide, I examine the ways in which labours of change extend to a field of sonically working efforts. Specifically, I study sites of activist involvement through which sounded actions perform the force of revolt against what is broadly defined as the Russian regime in its predatory formations and ecologies of harm.
The argument is laid out through a set of four operational modalities differentiated as the offender, the carer, the magician, and the prophet to approach sounds in their ways of performing political participation. I gather evidence from a range of case studies that share an activist stance and space of ethics, and that reveal an existing arena of insurrectionary activity. Each case gives expression to the experience of crises directly linked to Russia’s authoritarian, imperial, and capitalist politics and to the potency of the contingent moment of sounded resistances committed to causes and consequences of these politics. Spanning the spaces of activist presence across Ukraine, Uzbekistan, and Russia, the dissertation makes a claim to an ecology of sonic deeds, a politics of participation by which to collaborate and craft a form of caring relation
Performance Measurement in Supply Chain Greenhouse Gas Accounting: Assessing the State of the Practice
Get PDFSupply chain (upstream scope 3) greenhouse gas emissions often account for the majority of a company's total greenhouse gas emissions; however, they are often more challenging to measure and manage than emissions from direct sources (scope 1) or purchased energy (scope 2). Companies use various calculation methodologies and data sources to estimate supply chain emissions. This study evaluated manufacturing and services industry companies' supply chain emissions disclosures to CDP through the lens of GHG Protocol accounting and Supply Chain Performance Measurement (SCPM) principles. A set of indicators for assessing accounting practices according to three GHG Protocol accounting principles –measurement should be complete, accurate, and transparent– and two SCPM metric elements –metrics should show what is happening in a supply chain and be verifiable– were proposed. The results showed that, at most, 32% of disclosures follow both SCPM elements and GHG Protocol principles, and at most, emissions for 6% of supply chain categories are calculated using the emission data collected from suppliers. The results also showed that at least 39% and 32% of supply chain categories have transparency and completeness issues, respectively. This study also found that supply chain disclosures that are more transparent are more likely to be verified by a third party than those that are less transparent. However, a significant number of disclosures that lack transparency is still verified. Overall, the current completeness, accuracy, and transparency issues likely impede companies from reaching scope 3 emissions calculating goals