Today's data-driven applications, such as big data analytics, neural networks, and machine learning, require huge memory and computing resources. The latency and energy incurred for data movement between processing unit and memory has become a significant limitation (known as the "memory wall") as traditional techniques such as caching are no longer effective for the data-intensive applications that dominate modern computing. Solving the "memory wall" problem requires future computer technology to directly integrate memory and transistor devices with high density vertical interconnect accesses (vias) to provide parallel, high bandwidth memory access. However, the high temperature of silicon processing is not compatible with this 3D monolithic integration. A desired low temperature integration can be achieved by using two-dimensional (2D) materials which have intrinsically layered and flat atomic structures, because devices made of 2D materials can be fabricated at lower temperatures. This thesis introduces three exploratory studies on advancing 2D materials for developing 3D monolithic integrated systems. First, I focused on improving transistor performance and studied high-mobility 2D material black phosphorous (BP) transistors using various metals as the contact metal. This work achieved unipolar n-type BP transistors by using ultra-low work function metals, demonstrating record high n-type current. Furthermore, the study revealed the physical mechanisms of controlling doping and de-pinning effects for n- and p- type BP transistors. Beyond transistor studies, I then demonstrated the first 3D sequential monolithic integrated two levels of 1-transistor-1-resistor (1T1R) memory cells. The cell is fabricated entirely using 2D materials: hexagonal-Boron Nitride (hBN) serves as the resistive switching memory cell and monolayer Molybdenum Disulfide (MoS2) serves as the channel material for the transistor selector. In order to achieve a high-density memory array, a two-terminal selector with a high nonlinear current-voltage characteristic is necessary. I developed a new two-terminal selector utilizing a 2D material heterostructure with an H-shape energy barrier. This new device design in theory should have high on-state current density and virtually unlimited endurance due to its quantum tunneling mechanism. Our results characterized the first out-of-plane current through an ultra-thin (3 monolayer) heterojunction and provide the critical foundation for a high endurance selector using a 2D heterojunction. Due to the transferable feature of 2D materials, these processes can be easily adopted for 3D monolithic integrated transistors and memories in multiple logic and memory layers connected vertically by fine-grained nanoscale vias, thereby overcoming the "memory wall.".

In most higher education systems, the practice of a university hiring its own alumni is rarely dealt with. Commonly referred to as "endogamy" or "academic inbreeding", this has become part of some countries' academic culture, while others have proactively adopted policies to control for it and now experience large scholar mobility across institutions. Whether professors end up working for their alma mater for personal preference, structural limitations, or both, policies are better suitably contemplated and enforced once discrepancies in the academic job market are properly understood. Aiming at comprehending the Brazilian higher education academic job market, chapter 2 analyzes a large longitudinal database of active professors linked to graduate programs in Brazil. It describes not only the degree of academic endogamy in the system, but also the institutional and geographical characteristics which are associated with it. By looking within the faculty body of each university, a national ranking of institutions is established based on the highest degrees of endogamy. Finally, a regression model is adopted to identify characteristics that associate with such practice. Although at only 23% nationwide, results show the rate of endogamy significantly differs among distinct types of institutions, fields of knowledge, states, and university ranking tiers. This is especially true when analyzing elite research universities where such practice can amount to up to 70% of the faculty body. This analysis provides further evidence that endogamy is most likely to be found in established elite research institutions. These also happen to be the main providers and most prestigious consumers of the well-trained local academic workforce, supporting the conclusion that both internal and external factors could be influencing the high concentration of alumni scholars at prominent alma maters. Considering the shared characteristics of higher education systems, this trend ought to be similar in other young and developing systems worldwide. For decades, researchers have explored why academic endogamy happens and how it affects scientific production. Although consensus remains elusive, some studies associated endogamy with lower academic performance. Alternatively, external exposure seems to be correlated with higher research quality. Consequently, when measuring the effects of endogamy, authors should account for academic inertia as well as the different types of mobility. In chapter 3, a large longitudinal database on the Brazilian higher education system is used to measure the effects of endogamy and mobility on scientific production. In particular, this chapter examines whether types of scholars--such as those who never leave their doctoral institution ("Homegrown") or those that travel abroad to return later ("Adapted")--show differential results. Furthermore, it measures the changes in scholars' productivity when controlling for the distinct types of academic mobility throughout their careers. Results suggest that information about how mobile an academic before and during its career is relevant when estimating the effects of endogamy on academic productivity. Another interesting finding is that international academic experiences are correlated with higher likelihood of publishing at more prestigious journals. As shown in chapter 2, Research universities in Brazil commonly employ alumni as faculty members. Previous studies have measured the effects of endogamy on scientific production. Nevertheless, little is known about how it influences research collaboration. An environment with high academic endogamy would be expected to be rich in shared characteristics, or homophily. Consequently, these networks could be less welcoming to scholars trained at other universities. Such behavior can hurt opportunities of establishing ties with non-redundant contacts and limit the construction of new ideas. Chapter 4 collected data on research collaboration among 5,230 scholars in the University of São Paulo between 2000 and 2019 to understand how a network with high academic endogamy is structured, to identify if academic collaboration is more commonly found among those who share endogamy status, and to analyze if the likelihood of tie formation is distinct among native and non-native scholars. Results show increased collaboration over time. However, for both native and non-native scholars, ties are more likely between people with the same endogamy status. Furthermore, over time non-native scholars become less heterophilous, suggesting this institution could be missing out on opportunities of exploring non-redundant scholarship among its own faculty members

Mechanical properties are excellent predictors for device resilience to environmental stressors that accelerate the evolution of internal defects and cause delamination and device failure in layered structures. The fundamental mechanical and material properties of perovskite photovoltaics (PV) are first investigated to gain insight on relevant degradation modes and failure mechanisms. Perovskites are identified as the most mechanically fragile solar technology, further complicating the prospect of commercial viability. The mechanical integrity and residual film stresses of perovskite materials are connected with device stability, and strategies are developed to improve the thermomechanical reliability of perovskite PV in order to inform design criteria for stable, large-area solar modules. This work demonstrates the first industrially relevant attempt to address both scalable and fast open-air PV module manufacturing for the perovskite layer at production speeds that are > 10 m/min for the perovskite absorber layer. By implementing demonstrated open-air, rapid spray processing techniques with improved thermomechanical reliability of the perovskite absorber material at the highest reported throughput of any solar technology, significant steps are made towards reductions in manufacturing costs necessary to provide a new low-cost PV process that can compete with incumbent Si-based PV

Magnetic resonance imaging (MRI) is an important medical imaging modality for imaging soft tissue. A fundamental limitation of MRI is that there are tradeoffs between scan time, image resolution, and image signal-to-noise ratio (SNR). Typically, image resolution and SNR are not sacrificed because the images need to be diagnostically useful. Therefore, scan time is typically increased. For magnetic resonance angiography (MRA), this problem is exacerbated. Blood vessels are very small and, because of the ballistics of the heart pumping, the arteries are constantly in motion. Therefore, high resolution is required to resolve the vessels and the images need high SNR such that the vessels are distinguishable from noise. Additionally, while most clinical MRI acquires 2D slices, MRA typically requires 3D imaging because blood vessels typically do not lie in a single plane, which further extends scan time. To accelerate 3D MRA, we approached the problem from several directions. One method was to reduce the signal from all of the other tissue that we did not wish to image. We proposed a combined T2-preparation, outer volume suppression (OVS), fat saturation pulse to improve 3D coronary artery imaging. The pulse sequence consisted of the following pulse sequence: a 90°_-60 180°_{60} composite nonselective tip-down pulse, two 180°_Y hard pulses for refocusing, and a -90° spectral-spatial sinc tip-up pulse. The pulse sequence was played twice, first selective in the x-axis and then selective in the y-axis for a 2D OVS. Compared to another published T2-preparation OVS pulse sequence, the proposed pulse had superior image edge profile acutance values for the right (P< 0.05) and left (P< 0.05) coronary arteries, suggesting superior vessel sharpness. The proposed sequence also had superior SNR (P< 0.05) and passband-to-stopband ratio (P< 0.05). Reader scores and reader preference indicated superior image quality of the proposed sequence for both the right (P< 0.05) and left (P< 0.05) coronary arteries. Therefore, the proposed sequence had superior image quality and suppression, potentially enabling scan acceleration with reduced field of view imaging. Another technique to accelerate MRA was to enable more efficient acquisition trajectories. Non-Cartesian trajectories have the potential to be highly efficient but their efficiency is limited in practice by off-resonance. A residual convolutional neural network was trained to correct off-resonance artifacts in pediatric MRA exams (Off-ResNet). Training data was acquired from exams with a short readout scan (1.18 ms ± 0.38) and a long readout scan (3.35 ms ± 0.74) at 3 T. Short readout scans with longer scan times but negligible off-resonance blurring were used as reference images and augmented with additional off-resonance for supervised training examples. Long readout scans with greater off-resonance artifacts but shorter scan time were corrected by autofocus and Off-ResNet and compared with short readout scans by normalized root-mean-square error (NRMSE), structural similarity index (SSIM), and peak SNR (PSNR). Scans were also compared by scoring on eight anatomical features by two radiologists. Off-ResNet had superior NRMSE, SSIM, and PSNR compared to uncorrected images across ±1 kHz off-resonance (P< 0.01). Off-ResNet also had superior NRMSE over -677 Hz to +1 kHz and superior SSIM and PSNR over ±1 kHz compared with autofocus (P< 0.01). Radiologic scoring demonstrated that long readout scans corrected with Off-ResNet were noninferior to short readout scans (P< 0.05). The long readout scans were 59.3% shorter than the short readout scans, and Off-ResNet was able to correct the long readout scans to a noninferior quality compared to the diagnostically standard short readout scans. One limitation of Off-ResNet was that it used zero-order field maps to generate training data. While zero-order field maps are not realistic, it is also expensive to acquire large datasets with field maps. A generative adversarial network (GAN) was trained to generate realistic but not quantitatively accurate field maps for a given anatomical input image. The input to the generator network was an anatomical image from a single echo of a multi-echo T2*-IDEAL scan and the real examples of the corresponding field maps were also from the same set of T2*-IDEAL scans. We also proposed using a discriminator with Off-ResNet to improve off-resonance correction (Off-ResGAN). To evaluate the contributions of using a dataset with generated field maps and a GAN for off-resonance correction respectively, we trained four networks: 1. Off-ResNet and a dataset created with zero-order field maps, 2. Off-ResNet and a dataset created with generated field maps, 3. Off-ResGAN and a dataset created with zero-order field maps, and 4. Off-ResGAN and a dataset created with generated field maps. The four networks were evaluated on validation datasets created with zero-order field maps and generated field maps. The Off-ResGAN networks had superior NRMSE, SSIM, and PSNR on both datasets. With respect to datasets, each network performed best on the validation dataset that matched its training dataset. However, the networks trained with zero-order field maps tended to have worse performance on the generated field map validation dataset relative to the the performance of networks trained with generated field maps on the zero-order field map validation dataset. These initial results suggested that Off-ResGAN has superior image quality, and the benefit of using a dataset with generated field maps is increased robustness with respect to realistic off-resonant images. Undersampling is an important technique to accelerate scans. Undersampling reconstruction can involve a combination of parallel imaging and compressed sensing. An open question however is an optimal undersampling trajectory as measured by an image quality metric and with respect to a reconstruction technique. Moreover, as the scan proceeds, more information is collected and therefore increasingly better decisions can be made about the optimal undersampling trajectory. We proposed an approach using deep reinforcement learning to train an agent to learn from previous scans and combine the previously learned statistics with the new incoming data to determine the trajectory of the next readout. To frame the reinforcement learning problem, the environment was the reconstruction technique, the reward was the image metric, and the agent was a deep Q-learning convolutional neural network. Initial results with an L2 image quality metric and compressed sensing and deep learning reconstruction techniques indicated that the agent was capable of learning optimal trajectories. L2 was a special metric where we could analytically calculate the optimal trajectory. This suggested that the agent could learn optimal trajectories for other image metrics and reconstruction techniques where computing the optimal trajectory was intractable. In conclusion, the proposed techniques have demonstrated several strategies for accelerating 3D MRA. Through a combination of them, 3D MRA may become more feasible and accessible as a clinical tool

This dissertation comprises two parts. The first part studies acceptance rules on Bayesian statistical models and atomless probability spaces. We show that, for a very wide class of probabilistic learning problems, Leitgeb's probabilistic stability rule yields a notion of acceptance that either fails to be conjunctive (accepted hypotheses are not closed under finite conjunctions) or is trivial (only hypotheses with probability one are accepted). These results apply to most canonical Bayesian models involving exchangeable random variables, such as parametric models with Dirichlet priors over discrete distributions (and, in particular, to every confirmation function in Carnap's gamma-delta continuum). These results exhibit a serious tension for the stability rule: in Bayesian statistical models, important properties of priors that are conducive to inductive learning—Bayesian consistency, as well as certain symmetries in the agent's probability assignments—act against conjunctive acceptance. We then investigate impossibility theorems for acceptance rules on atomless spaces. We first strengthen a theorem by Smith, which generalises a form of the Lottery Paradox to atomless probability spaces, revealing an important conflict between certain invariance constraints on rational acceptance and logical closure requirements on accepted hypotheses. We introduce a refined class of acceptance rules, under which acceptance is sensitive to the structure of the evidence available (for instance, the acceptance of a hypothesis can depend on the structure of the observable data in the given experimental setup). While these rules escape Smith's result, we prove that, on learning problems based on standard Borel probability spaces, sufficiently invariant acceptance rules are either (1) not conjunctive, (2) inconsistent, or (3) trivial in a learning-theoretic (topological) sense. These results raise a number of questions about the context-sensitivity of acceptance, the aggregation of uncertain information in highly symmetric probability models, and the relationship between acceptance rules and statistical testing procedures. In the second part, we study the modal logics of approximation operators on Boolean algebras. We axiomatise the logics of approximation operators generated by arbitrary sets, sublattices, complete sublattices and subalgebras. For each class of substructures, we provide a representation theorem which characterises the corresponding class of closure-kernel algebras, and prove a completeness theorem, identifying the corresponding bimodal logic. Our investigations show that these approximation semantics allow to recover a natural family of modal systems: the fusions S4+S4 and EMNT4+EMNT4, the temporal logic S4t, as well as S4 and S5 are all shown to be logics of approximation operators. By varying the structural constraints imposed on the set D which generates the approximation operators, we obtain a natural family of modal logics. The logic of approximation operators generated by arbitrary subsets is the subnormal logic EMNT4+EMNT4. We characterise the corresponding class of algebras: that is, algebras with closure and kernel operators representable as approximation operators of the same underlying set. We show that representable algebras do not constitute an algebraic variety. We then prove that the complete logic of sublattice-generated closures and kernels is the fusion S4+S4. Our completeness proof, which uses a result by van Benthem et al. on vertical-horizontal bitopologies on the rational plane, relies on showing that the interior-closure algebras of pairwise zero-dimensional bitopological spaces correspond to sublattice-generated closure-kernel algebras. We note, as a corollary, that S4+S4 is the bimodal logic of pairwise zero-dimensional bitopological spaces. When D is a complete sublattice, we obtain exactly the temporal logic S4t, exploiting a Tarski-like duality between complete sublattices and pairs of closure and kernel operators. When D is a subalgebra, the two modalities collapse into one as they become each other's duals, and we obtain monomodal S5 (for complete subalgebras) and S4 (for subalgebras in general)

It is widely recognized that physical activity is critical for individuals' health and longevity. Tremendous resources are being devoted to initiatives aiming to increase people's physical activity levels, including public health campaigns, workplace health programs, and the development of health technology. However, this focus on physical activity behavior has obscured another, related factor that may powerfully shape health and wellbeing----that is, mindsets about physical activity. This dissertation shines a light on individuals' mindsets about the adequacy and health consequences of their level of physical activity (activity adequacy mindsets). It investigates the causes, consequences, and mechanisms of action of these mindsets, and it develops interventions to help individuals adopt beneficial mindsets and thereby improve their health and wellbeing. In Chapter I, I introduce the entire dissertation. I review the literature on mindsets and health, and introduce the novel construct of activity adequacy mindset. I then describe and provide context for five open questions that are critical to understanding and harnessing activity adequacy mindsets. First, do activity adequacy mindsets causally influence health and wellbeing, independently of actual activity? Second, do these mindsets have long-term consequences? Third, what are the real-world sources of activity adequacy mindsets? Fourth, what are the processes explaining how activity adequacy mindsets affect health and wellbeing? And fifth, how can we harness activity adequacy mindsets to promote people's health and wellbeing at scale? The following chapters address each of these open questions in turn. In Chapter II, I investigate whether activity adequacy mindsets may influence important long-term outcomes. Specifically, I examine whether individuals' perceived levels of physical activity compared to others----a proxy for activity adequacy mindset----predict mortality. Survival analysis of data from three nationally representative samples with a total sample size of 61,141 U.S. adults and follow-up periods of up to 21 years shows that perceived physical activity does indeed predict mortality, controlling for actual amounts of physical activity. In Chapter III, I test if physical activity guidelines may represent one important real-world source of activity adequacy mindsets. Two experimental studies show that guidelines prescribing a relatively high amount of physical activity (such as the current official guidelines) may unintentionally induce negative activity adequacy mindsets, and thereby undermine individuals' self-efficacy, physical activity behavior, and health. In Chapter IV, I present a longitudinal field experiment leveraging the sensing and feedback capabilities of Apple Watch to study the causal effects of activity adequacy mindsets on health and explore the processes that may explain how mindsets create these effects. Additionally, I develop and test a scalable meta-mindset intervention designed to empower individuals to consciously adopt more beneficial mindsets and thereby improve their own health and wellbeing

In the last decade, online learning platforms have dramatically increased access to educational materials, and replaced and supplemented traditional co-located instruction. Throughout this shift, instructors have struggled with how best to create and deploy materials to online platforms and understand their effectiveness when scaled across hundreds or thousands of learners. To address these complementary problems, this thesis introduces new systems and methods for both learning analytics and adaptive instruction. In OARS, we demonstrate a real-time learning analytics system deployed across more than ten online courses with tens of thousands of learners. We then report on the value of learning analytics from the perspective of course instructors. Learning from these perspectives, we next introduce a system for adaptively scheduling educational activities through reinforcement learning (RL). Without any skill labels, this model learns how to assign educational activities in a way that maximizes learning gains while minimizing redundant work. Finally, in a randomized controlled experiment, we show that our RL scheduling algorithm helped to improve the educational experience for online learners by reducing the number of learning activities required to produce the same learning gains

Adornos Arbeiten kreisen beständig um die Frage nach den Bedingungen der Möglichkeit eines besseren und gerechteren Lebens. Nimmt man das Glück als Ausgangspunkt von Adornos Denken ernst, so kommt der Kunst die Bedeutung zu, eine andere, gelungenere Welt aufscheinen zu lassen. Ausgehend von dieser Beobachtung untersucht die Autorin, wie Kunstwerke nach Adorno beschaffen sein müssen, um ein Glücksversprechen artikulieren zu können. Sie verbindet diese Frage der Produktion von Kunst mit einer rezeptionsästhetischen Perspektive, die das Glück der ästhetischen Erfahrung selbst in den Blick nimmt. Gerade in Adornos literatur- und kulturtheoretischen Essays, u.a. zu Beckett, Kafka, Hölderlin, Helms und Goethe, aber auch zur Heiterkeit der Kunst und zum Kitsch, finden sich Elemente einer Theorie des Glücks, so ihre These, die entscheidende Bedeutung für Adornos gesamtes Denken haben. Durch die historische Kontextualisierung seiner Essays rückt Groß das "Glück am Ästhetischen" in Adornos Werk ins Zentrum und zeigt, dass für ihn das Gelingen von Kunst einzig davon abhängt, ob ihr glücksversprechende Potentiale innewohnen