Aujourd’hui et à l’avenir, nous devrons interagir de plus en plus avec des représentations visuelles nous invitant à poser des gestes d’achat qui contribuent, par exemple, à la lutte contre les changements climatiques. Ainsi une autodéclaration, un écolabel ou une étiquette environnementale multicritères pourront nous inciter (ou non) à recycler un produit, à le choisir ou encore à le comparer avec d’autres produits en fonction de l’information que cette représentation et les signes qui la composent lui feront dire. Ce scénario repose sur l’idée selon laquelle la typographie, les formes, les couleurs et la configuration de cet affichage environnemental possèdent de multiples formes d’agentivité, soit la capacité d'agir, autrement dit, de faire une différence (Castor & Cooren, 2006), une capacité qu’il s’agit donc d’anticiper. La conception de ces représentations visuelles, en apparence inanimées, mais qui néanmoins sont censées faire agir le public auquel elles s’adressent, s’inscrit dans ma pratique professionnelle : le design graphique. En raison de l’information plus ou moins complexe qui doit être véhiculée, ces signes — des icônes, indices et symboles (selon la classification de Peirce) — devront être choisis et conçus par une équipe multidisciplinaire d’experts en anticipant les multiples formes d’agentivité qu’ils pourraient posséder une fois configurés. Or, bien que le processus de conception des représentations visuelles ait été abondamment étudié (Self & Goldschmidt, 2018), nous en savons encore peu sur la manière dont sont anticipées les différentes formes d’agentivité que ces représentations pourront/devront posséder en aval, au terme de la trajectoire. Les études en design suggèrent qu’une « vision professionnelle partagée » du projet à concevoir émerge à travers les représentations visuelles — des objets-frontière et intermédiaires (Vinck, 2009) — rencontrées en amont de la trajectoire (Comi, Jaradat, & Whyte, 2019). Dans ce contexte, je propose de répondre à la question suivante : Quelle est la trajectoire de conception collaborative et multidisciplinaire des multiples formes d’« agentivité anticipée » d’une représentation visuelle ?Pour répondre à cette question, je me suis inspirée de l’ethnométhodologie, (Garfinkel, 1984), de la sociologie des associations (Latour, 2006 ; Callon, 1986) et de l’approche constitutive de la communication organisationnelle (CCO, Schoeneborn et al., 2014), pour étudier les interactions entre les principaux actants — des éléments humains et autres qu’humains — qui agissent lors de la trajectoire de conception d’un affichage environnemental de la durée de vie d’un appareil électronique. La trajectoire en six temps et quatre mouvements (zooming-out et passé/ futur), que j’analyse à travers la plateforme de présentation Prezi®, oscille entre les moments clés de collaboration mobilisant les membres de notre équipe multidisciplinaire de conception et les séances de consultation menées avec un comité consultatif d’experts provenant également de différentes disciplines. Le regard réflexif que je pose sur la conception de « l’agentivité anticipée » des représentations visuelles par les représentations visuelles — première contribution de cette thèse — suggère que l’affichage environnemental s’est matérialisé principalement grâce aux interactions situées avec des actants hétérogènes (in)tangibles relevant du passé, du présent ou du futur. En effet, de multiples porte-parole parlant au nom, par exemple, du public (consommateur ou usager), du fabricant, du réparateur, des autorités politiques ou des (éco)concepteurs de produits se sont ainsi exprimés par l’entremise des représentations visuelles conçues tout au long de cette trajectoire. La configuration de ces signes — lettres, chiffres, formes, etc. — a provoqué de multiples (dés)accords générés par nos visions professionnelles et/ou personnelles respectives. Ces visions « (dé)partagées » de l’« agentivité anticipée » des représentations visuelles ont cependant grandement contribué au processus de conception en dévoilant des actants (in)visibles jusqu’alors ainsi que leurs relations de pouvoir. Cette « compréhension (dé)partagée » collaborative et multidisciplinaire lors du processus de conception — deuxième contribution de cette thèse — pourrait être une voie à suivre pour créer des dispositifs visuels mobilisant davantage les êtres humains et autres qu’humains vers une transition écologique réussie (Brullot et al., 2017). / Nowadays and in the future, we will interact more and more often with visual representations that will invite us to take actions that contribute, for example, to the fight against climate change. Thus, a self-declaration, an eco-label or a multi-criteria environmental label will encourage us (or not) to recycle a product, to choose it or to compare it with other products according to the information that this representation and the signs that constitutes it should convey. In this scenario, the typography, shapes, colors and configuration of this environmental label will possess multiple forms of agency, the capacity to act, that is, to make a difference (Castor & Cooren, 2006), a capacity that has to be anticipated. The design of these visual representations, apparently inanimate but which nevertheless communicate and make people do things, is part of my professional practice: graphic design. Because of the more or less complex information they convey, these signs — icons, index and symbols, according to Peirce’s classification — will often have to be designed by a multidisciplinary team of experts, anticipating the multiple forms of agency they might possess once configured. However, although the process of designing visual representations has been extensively studied, we still know little about the collaborative design of the different forms of agency these representations may (or should) possess at the end of the process. Design studies suggest that a “shared professional vision” of the project to be designed emerges through visual representations — boundary and intermediary objects (Vinck, 2009) — encountered early in the trajectory (Comi, Jaradat, & Whyte, 2019). In this context, I propose to address the following research question: What is the collaborative and multidisciplinary design trajectory of the multiple forms of “anticipated agency” of a visual representation? To answer this question, I drew on ethnomethodology, (Garfinkel, 1984), the sociology of associations (Latour, 2006; Callon, 1986) as well as the communicative constitution of organization approach (CCO, Schoeneborn et al., 2014) to document and analyze the interactions between key actants — humans and other-than-human — that are making a difference during the design trajectory of an environmental label conceived to inform about the life span of an electronic device. The six-beat and four-movement trajectory (zooming-out and past/ future), that I analyse through the Prezi® presentation online platform, oscillates between key moments of collaboration between members of our multidisciplinary design team and consultation sessions conducted with an advisory committee of experts also from different disciplines. My reflective vision through the design of the “anticipated agency” of visual representations with visual representations — the first contribution of this dissertation — suggests that the environmental label materialized primarily through situated interactions with heterogeneous (in)tangible actants belonging to the past, present, or future. Indeed, multiple spokespersons speaking on behalf of, for example, the public (the consumer or the user), the manufacturer, the repairer, the political authorities or the (eco)designers of products have thus expressed themselves through the visual representations designed. The configuration of these signs — letters, numbers, shapes, colors, etc. — has induced multiple (dis) agreements generated by our respective professional and/or personal visions. However, these (un)shared visions of the “anticipated agency” of those visual representations, however, greatly contributed to the design process by revealing previously unseen actants and their power relations. This collaborative and multidisciplinary “(un)shared understanding” of the design process — the second contribution of this dissertation — could be a way forward to create visual devices that further mobilize human and other-than-human toward a successful ecological transition (Brullot et al., 2017).

<p>Multifunctional additive manufacturing has the immense potential of addressing present needs within structural health monitoring by enabling a new additive manufacturing paradigm that redefines what a sensor is, or what sensors should resemble. To achieve this, the properties of printed components must be precisely tailored to meet structure specific and application specific requirements. However due to the limited number of commercially available multifunctional filaments, this research investigates the in-house creation of adaptable piezoresistive multifunctional filaments and their potential within structural health monitoring applications based upon their characterized piezoresistive responses. To do so, a rigid polylactic acid based-filament and a flexible thermoplastic polyurethane based-filament were modified to impart piezoresistive properties using carbon nanofibers. The filaments were produced using different mixing techniques, nanoparticle concentrations, and optimally selected manufacturing parameters from a design of experiments approach. The resulting filaments exhibited consistent resistivity values which were found to be less variable under specific mixing techniques than commercially available multifunctional filaments. This improved consistency was found to be a key factor which held back currently available piezoresistive filaments from fulfilling needs within structural health monitoring. To demonstrate the ability to meet these needs, the piezoresistive responses of three dog-bone shaped sensor sizes were measured under monotonic and cyclic loading conditions for the optimally manufactured filaments. The characterized piezoresistive responses demonstrated high strain sensitivities under both tensile and compressive loads. These piezoresistive sensors demonstrated the greatest sensitivity in tension, where all three sensor sizes exhibited gauge factors over 30. Cyclic loading supported these results and further demonstrated the accuracy and reliability of the printed sensors within SHM applications.</p>

<p>Convolutional neural networks (CNNs) have become important workloads due to their<br> impressive accuracy in tasks like image classification and recognition. Convolution operations<br> are compute intensive, and this cost profoundly increases with newer and better CNN models.<br> However, convolutions come with characteristics such as sparsity which can be exploited. In<br> this dissertation, we propose three different works to capture sparsity for faster performance<br> and reduced energy. </p> <p><br></p> <p>The first work is an accelerator design called <em>SparTen</em> for improving two-<br> sided sparsity (i.e, sparsity in both filters and feature maps) convolutions with fine-grained<br> sparsity. <em>SparTen</em> identifies efficient inner join as the key primitive for hardware acceleration<br> of sparse convolution. In addition, <em>SparTen</em> proposes load balancing schemes for higher<br> compute unit utilization. <em>SparTen</em> performs 4.7x, 1.8x and 3x better than dense architecture,<br> one-sided architecture and SCNN, the previous state of the art accelerator. The second work<br> <em>BARISTA</em> scales up SparTen (and SparTen like proposals) to large-scale implementation<br> with as many compute units as recent dense accelerators (e.g., Googles Tensor processing<br> unit) to achieve full speedups afforded by sparsity. However at such large scales, buffering,<br> on-chip bandwidth, and compute utilization are highly intertwined where optimizing for<br> one factor strains another and may invalidate some optimizations proposed in small-scale<br> implementations. <em>BARISTA</em> proposes novel techniques to balance the three factors in large-<br> scale accelerators. <em>BARISTA</em> performs 5.4x, 2.2x, 1.7x and 2.5x better than dense, one-<br> sided, naively scaled two-sided and an iso-area two-sided architecture, respectively. The last<br> work, <em>EUREKA</em> builds an efficient tensor core to execute dense, structured and unstructured<br> sparsity with losing efficiency. <em>EUREKA</em> achieves this by proposing novel techniques to<br> improve compute utilization by slightly tweaking operand stationarity. <em>EUREKA</em> achieves a<br> speedup of 5x, 2.5x, along with 3.2x and 1.7x energy reductions over Dense and structured<br> sparse execution respectively. <em>EUREKA</em> only incurs area and power overheads of 6% and<br> 11.5%, respectively, over Ampere</p>

<p>The "Maker Movement" is a cultural phenomena rooted in DIY culture, which stresses making devices and creations on your own rather than purchasing it ready-made. At the core of the Maker Movement, is the "Maker Mindset"; a collection of attitudes, beliefs and behaviors that emphasize the importance of creativity, experimentation and innovation in the learning process. Since the Maker Mindset embodies constructionist principles at its core that push makers to experiment and problem-solve by collaborating with fellow makers through hands-on activities, it can be said that these activities comprise of Creative, Learning and Collaborative experiences. While Internet-of-Things devices have long been used to enhance these activities, research pertaining to using Augmented Reality in tandem with IoT for the purpose of enhancing experiences core to the Maker Mindset is relatively unexplored. Three different systems were developed with the goal of addressing this -- MicrokARts, ShARed IoT and MechARspace. Each system focuses on enhancing one of the three core experiences through AR-compatible IoT devices, whilst ensuring that they do not require prerequisite knowledge in order to author AR experiences. These systems were evaluated through user studies and testing over a variety of age-groups, with each system successfully enhancing one core experience each through the use of AR-IoT interactions.</p>